OPL Language Reference Manual

OPL_Language_Reference_Manual

User Manual:

Open the PDF directly: View PDF PDF.
Page Count: 152 [warning: Documents this large are best viewed by clicking the View PDF Link!]

IBM ILOG CPLEX Optimization Studio
OPL Language Reference Manual
Ve r s i o n 1 2 R e l e a s e 6

IBM ILOG CPLEX Optimization Studio
OPL Language Reference Manual
Ve r s i o n 1 2 R e l e a s e 6

Copyright notice
Describes general use restrictions and trademarks related to this document and the software described in this
document.
© Copyright IBM Corp. 1987, 2015
US Government Users Restricted Rights - Use, duplication or disclosure restricted by GSA ADP Schedule Contract
with IBM Corp.
Trademarks
IBM, the IBM logo, and ibm.com are trademarks or registered trademarks of International Business Machines Corp.,
registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other
companies. A current list of IBM trademarks is available on the Web at "Copyright and trademark information" at
www.ibm.com/legal/copytrade.shtml.
Adobe, the Adobe logo, PostScript, and the PostScript logo are either registered trademarks or trademarks of Adobe
Systems Incorporated in the United States, and/or other countries.
Linux is a registered trademark of Linus Torvalds in the United States, other countries, or both.
UNIX is a registered trademark of The Open Group in the United States and other countries.
Microsoft, Windows, Windows NT, and the Windows logo are trademarks of Microsoft Corporation in the United
States, other countries, or both.
Java and all Java-based trademarks and logos are trademarks or registered trademarks of Oracle and/or its
affiliates.
Other company, product, or service names may be trademarks or service marks of others.
© Copyright IBM Corporation 1987, 2015.
US Government Users Restricted Rights – Use, duplication or disclosure restricted by GSA ADP Schedule Contract
with IBM Corp.
Contents
Figures ...............v
Tables ...............vii
Language Reference Manual .....ix
Chapter 1. Why an Optimization
Programming Language? .......1
Chapter 2. OPL, the modeling language 3
Building a model .............3
Data types ...............4
Basic data types ............4
Data structures ............7
Data sources ..............15
Data initialization ...........15
Database initialization ..........24
Spreadsheet Input/Output ........29
Data consistency............32
Preprocessing of data ..........34
Pitfall of lazy initialization of data ......36
Decision types .............37
Decision variables ...........37
Expressions of decision variables ......38
Dynamic collection of elements into arrays . . . 39
Expressions ..............41
Usage of expressions ..........41
Data and decision variable identifiers.....41
Integer and float expressions .......42
Aggregate operators ..........43
Piecewise-linear functions ........43
Discontinuous piecewise-linear functions . . . 45
Set expressions ............48
Boolean expressions ..........50
Constraints ..............51
Introduction .............51
Using constraints ...........51
Constraint labels............53
Types of constraints ..........58
Limitations on constraints ........64
Formal parameters ............64
Basic formal parameters .........64
Tuples of parameters ..........66
Filtering in tuples of parameters ......67
Scheduling ..............68
Introduction .............68
Piecewise linear and stepwise functions ....69
Interval variables ...........71
Unary constraints on interval variables ....74
Precedence constraints between interval variables 75
Constraints on groups of interval variables . . . 76
A logical constraint between interval variables:
presenceOf .............76
Expressions on interval variables ......77
Sequencing of interval variables ......78
Cumulative functions ..........82
State functions ............86
Notation ..............89
Chapter 3. IBM ILOG Script for OPL . . 91
Language structure ............91
Syntax ...............91
Expressions in IBM ILOG Script ......92
Statements..............98
Built-in values and functions ........103
Numbers ..............103
IBM ILOG Script strings .........110
IBM ILOG Script Booleans ........117
IBM ILOG Script arrays .........119
Objects ..............124
Dates ...............126
The null value ............130
The undefined value ..........130
IBM ILOG Script functions ........131
Miscellaneous functions .........132
Index ...............135
© Copyright IBM Corp. 1987, 2015 iii
iv OPL Language Reference Manual
Figures
1. A Piecewise-linear function. .......44
2. The discontinuous piecewise-linear function
sign() ..............46
3. Discontinuous costs ..........48
4. Labeled assert............56
© Copyright IBM Corp. 1987, 2015 v
vi OPL Language Reference Manual
Tables
1. Escape sequences inside literal strings ....5
2. Operations allowed on sets .......12
3. Initializing sets in the model file .....22
4. Database connection strings .......25
5. A discontinuous cost function ......47
6. Functions over sets .........49
7. Reference syntax of script variables ....93
8. Property access syntax .........94
9. Assignment operator syntax .......94
10. Syntactic shorthand ..........95
11. Function call syntax..........95
12. Special keyword syntax ........96
13. Special operator syntax ........96
14. Other operator syntax .........98
15. Conditional statement syntax ......99
16. Loop syntax ............99
17. Declaration syntax for script variables 101
18. Function definition syntax .......102
19. Default value syntax .........103
20. Number method ..........106
21. Numeric functions ..........106
22. Numeric constants ..........107
23. Numeric operators ..........108
24. Escape sequences in strings .......110
25. Examples of string literals using escape
sequences .............111
26. String property ...........112
27. String methods ...........112
28. String functions ...........114
29. String operators ...........116
30. Boolean method...........118
31. Logical operators ..........118
32. Array constructor ..........120
33. Array properties ..........121
34. Array methods ...........122
35. Object constructor ..........125
36. User-defined constructors .......125
37. Built-in method ...........125
38. Date constructor ..........126
39. Date methods ...........128
40. Date functions ...........129
41. Methods of null...........130
42. Methods of undefined ........131
43. Function methods ..........131
44. Miscellaneous functions ........132
© Copyright IBM Corp. 1987, 2015 vii
viii OPL Language Reference Manual
Language Reference Manual
This manual provides reference information about IBM®ILOG®Optimization
Programming Language (OPL), the modeling language used in CPLEX®Studio.
For details of prerequisites, naming conventions, and other general information,
see How to use the documentation in the Starting Kit.
© Copyright IBM Corp. 1987, 2015 ix
xOPL Language Reference Manual
Chapter 1. Why an Optimization Programming Language?
OPL is a modeling language for combinatorial optimization that aims at
simplifying the solving of optimization problems.
Linear programming, integer programming, and combinatorial optimization
problems arise in a variety of application areas, which include planning,
scheduling, sequencing, resource allocation, design, and configuration.
In this context, OPL is a modeling language for combinatorial optimization that
aims at simplifying the solving of these optimization problems. As such, it
provides support in the form of computer equivalents for modeling linear,
quadratic, and integer programs, and provides access to state-of-the-art algorithms
for linear programming, mathematical integer programming, and quadratic
programming.
Within the IBM ILOG OPL product, OPL as a modeling language has been
redesigned to better accommodate IBM ILOG Script, its associated script language.
© Copyright IBM Corp. 1987, 2015 1
2OPL Language Reference Manual
Chapter 2. OPL, the modeling language
Presents the modeling language of IBM ILOG OPL, namely: the overall structure of
OPL models; the basic modeling concepts; how data can be initialized internally as
it is declared or externally in a .dat file; how to connect to, read from, and write to
databases and spreadsheets; expressions and relations; constraints; and formal
parameters.
Building a model
Describes the primary elements of the OPL language and how they are used to
build an optimization model.
The basic building blocks of OPL are integers, floating-point numbers, identifiers,
strings, and the keywords of the language. Identifiers in OPL start with a letter or
the underscore character(_)andcancontain only letters, digits, and the
underscore character. Note that letters in OPL are case-sensitive. Integers are
sequences of digits, possibly prefixed by a minus sign. Floats can be described in
decimal notation (3.4 or -2.5) or in scientific notation (3.5e-3 or -3.4e10).
The OPL reserved words are listed in OPL keywords.
Comments in OPL are written in between /* and */ as in:
/*
This is a
multiline comment
*/
The characters // also start a comment that terminates at the end of the line on
which they occur as in:
dvar int cost in 0..maxCost; // decision variable
An OPL model consists of:
va sequence of declarations
voptional preprocessing instructions
vthe model/problem definition
voptional postprocessing instructions
voptional flow control (main block)
Details about these elements are given in the chapters that follow.
A simple model (volsay.mod)
A typical model is shown here.
dvar float+ Gas;
dvar float+ Chloride;
maximize
40 * Gas + 50 * Chloride;
subject to {
ctMaxTotal:
Gas + Chloride <= 50;
© Copyright IBM Corp. 1987, 2015 3
ctMaxTotal2:
3*Gas+4*Chloride <= 180;
ctMaxChloride:
Chloride <= 40;
}
In this example, there is only the declarative initial part and the model definition.
There is no preprocessing, postprocessing, or flow control.
Data types
Describes basic data types and data structures available for modeling data in OPL.
Basic data types
Describes integers, floats, strings, piecewise linear functions, and stepwise
functions in OPL.
Integers
Describes integers (int) in OPL.
Shows how to declare integers in the OPL language.
OPL contains the integer constant maxint, which represents the largest positive
integer available. OPL provides the subset of the integers ranging from -maxint to
maxint as a basic data type.
A declaration of the form
int i = 25;
declares an integer, i, whose value is 25.
The initialization of an integer can be specified by an expression. For instance, the
declaration
intn=3;
int size = n*n;
initializes size as the square of n. Expressions are covered in detail in
“Expressions” on page 41.
Floats
Describes floats (float) in OPL.
Shows how to declare floats in the OPL language.
OPL also provides a subset of the real numbers as the basic data type float. The
implementation of floats in OPL obeys the IEEE 754 standard for floating-point
arithmetic and the data type float in OPL uses double-precision floating-point
numbers. OPL also has a predefined float constant, infinity, to represent the IEEE
infinity symbol. Declarations of floats are similar to declarations of integers.
The declaration
float f = 3.2;
declares a float fwhose value is 3.2.
The value of the float can be specified by an arbitrary expression.
4OPL Language Reference Manual
Strings
Describes strings (string) in OPL.
Shows how to declare strings in the OPL language.
OPL supports a string data type. The excerpt
{string} Tasks = {"
masonry","carpentry","plumbing","ceiling",
"roofing","painting","windows","facade",
"garden","moving"};
defines and initializes a set of strings. Strings can appear in tuples and index
arrays.
Strings are manipulated in the preprocessing phase via ILOG Script. See Chapter 3,
“IBM ILOG Script for OPL,” on page 91 for details on the scripting language.
The OPL parser supports the following escape sequences inside literal strings:
Table 1. Escape sequences inside literal strings
\b backspace
\t tab
\n newline
\f form feed
\r carriage return
\" double quote
\\ backslash
\ooo octal character ooo
\xXX hexadecimal character XX
To continue a literal string over several lines, you need to escape the new line
character:
"first line \
second line"
Piecewise linear functions
Describes piecewise linear functions (pwlFunction) in OPL.
Shows how to declare piecewise linear functions in the OPL language.
Piecewise linear functions are typically used to model a known function of time,
for instance the cost incurred for completing an activity after a known date.
Note that you must ensure that the array of values T[i] is sorted.
The meanings of the S,T, and Vvectors are described in “Piecewise linear and
stepwise functions” on page 69, in the Language Reference Manual.
Syntax
pwlFunction F = piecewise(i in 1..n){ S[i]->T[i]; S[n+1] } (t0, v0);
pwlFunction F = piecewise{ V[1]->T[1], ..., V[n]->T[n], V[n+1] };
pwlFunction F[i in ...] = piecewise (...)[ ... ];
Chapter 2. OPL, the modeling language 5
Example
int n=2;
float objectiveforxequals0=300;
float breakpoint[1..n]=[100,200];
float slope[1..n+1]=[1,2,-3];
int points[1..n+1] = [5, 150, 400];
float values[x in 1..n+1] = piecewise(i in 1..n) {slope[i] -> breakpoint[i]; slope[n+1]}(0,0) points[x];
execute {
writeln( values );
}
dvar int x;
maximize piecewise(i in 1..n) {slope[i] -> breakpoint[i]; slope[n+1]}(0,objectiveforxequals0) x;
subject to {true;}
Piecewise linear functions are covered in detail in “Piecewise linear and stepwise
functions” on page 69.
Stepwise functions
Describes stepwise functions (stepFunction) in OPL.
Shows how to declare stepwise functions in the OPL language.
Stepwise linear functions are typically used to model the efficiency of a resource
over time. A stepwise function is a special case of piecewise linear function where
all slopes are equal to 0and the domain and image of Fare integer.
Note that you must ensure that the array of values T[i] is sorted.
Syntax
stepFunction F = stepwise(i in 1..n){ V[i]->T[i]; V[n+1] };
stepFunction F = stepwise{ V[1]->T[1], ..., V[n]->T[n], V[n+1] };
stepFunction F[i in ...] = stepwise (...)[ ... ];
Example
A declaration of the form
stepFunction f=stepwise {0->3; 2};
assert f(-1)==0;
assert f(3)==2;
assert f(3.1)==2;
declares a stepwise function, f.
Example
Another example, declaring the stepwise function F2:
stepFunction F2 = stepwise{ 0->0; 100->20; 60->30; 100 };
int ii= F2( 10 );
6OPL Language Reference Manual
execute {
writeln( ii );
writeln( F2( 25 ) );
}
Stepwise functions are covered in detail in “Piecewise linear and stepwise
functions” on page 69.
Data structures
Describes how the basic data types can be combined using arrays, tuples, and sets
to obtain complex data structures.
Ranges
Integer ranges are fundamental in OPL, since they are often used in arrays and
variable declarations, as well as in aggregate operators, queries, and quantifiers.
Declaring ranges
To specify an integer range, you give its lower and upper bounds, as in
range Rows = 1..10;
which declares the range value 1..10. The lower and upper bounds can also be
given by expressions, as in
intn=8;
range Rows = n+1..2*n+1;
Once a range has been defined, you can use it as an array indexer.
Whenever a range is empty, i.e. its upper bound is less than its lower bound, it is
automatically normalized to 0..-1 (in other words, all empty ranges are equal).
The range declaration
An integer range is typically used:
vas an array index in an array initialization expression
range R = 1..100;
int A[R]; // A is an array of 100 integers
vas an iteration range
range R = 1..100;
forall(i in R) {
//element of a loop
...
}
vas the domain of an integer decision variable
dvar int i in R;
The range float declaration
Arange float data type consists of a couple of float values specifying an interval.
It is typically used as the domain of a floating-point decision variable.
range float X = 1.0..100.0;
dvar float x in X;
Chapter 2. OPL, the modeling language 7
Arrays
Describes one-dimensional arrays and multidimensional arrays.
Arrays are fundamental in many applications.
One-dimensional arrays
One-dimensional arrays are the simplest arrays in OPL and vary according to the
type of their elements and index sets. A declaration of the form
int a[1..4] = [10, 20, 30, 40];
declares an array of four integers a[1],...,a[4] whose values are 10,20,30, and
40. It is of course possible to define arrays of other basic types. For instance, the
instructions
int a[1..4] = [10, 20, 30, 40];
float f[1..4] = [1.2, 2.3, 3.4, 4.5];
string d[1..2] = [“Monday”, “Wednesday”];
declare arrays of natural numbers, floats, and strings, respectively.
The index sets of arrays in OPL are very general and can be integer ranges and
arbitrary finite sets. In the examples so far, index sets were given explicitly, but it is
possible to use a previously defined range, as in
range R = 1..4;
int a[R] = [10, 20, 30, 40];
The declaration:
int a[Days] = [10, 20, 30, 40, 50, 60, 70];
describes an array indexed by a set of strings; its elements are
a[“Monday”],...,a[“Sunday”].
Arrays can also be indexed by finite sets of arbitrary types. This feature is
fundamental in OPL to exploit sparsity in large linear programming applications,
as discussed in detail in Exploiting sparsity in the Language User’s Manual.
For example, the declaration:
tuple Edges {
int orig;
int dest;
}
{Edge} Edges = {<1,2>, <1,4>, <1,5>};
int a[Edges] = [10,20,30];
defines an integer array, a, indexed by a finite set of tuples. Its elements are
a[<1,2>], a[<1,4>], and a[<1,5>]. Tuples are described in detail in “Tuples” on
page 9.
Multidimensional arrays
OPL supports the declaration of multidimensional arrays (see “Data initialization”
on page 15 about the ellipsis syntax). For example, the declaration:
int a[1..2][1..3] = ...;
declares a two-dimensional array, a, indexed by two integer ranges. Indexed sets of
different types can of course be combined, as in
8OPL Language Reference Manual
int a[Days][1..3] = ...;
which is a two-dimensional array whose elements are of the form a[Monday][1].It
is interesting to contrast multidimensional and one-dimensional arrays of tuples.
Consider the declaration:
{string} Warehouses = ...;
{string} Customers = ...;
int transp[Warehouses,Customers] = ...;
that declares a two-dimensional array transp. This array may represent the units
shipped from a warehouse to a customer. In large-scale applications, it is likely that
a given warehouse delivers only to a subset of the customers. The array transp is
thus likely to be sparse, i.e. it will contain many zero values.
The sparsity can be exploited by declarations of the form:
{string} Warehouses = ...;
{string} Customers = ...;
tuple Route {
string w;
string c;
}
{Route} routes = ...;
int transp[routes] = ... ;
This declaration specifies a set, routes, that contains only the relevant pairs
(warehouse, customer). The array transp can then be indexed by this set,
exploiting the sparsity present in the application. It should be clear that, for
large-scale applications, this approach leads to substantial reductions in memory
consumption.
You can initialize arrays by listing its values, as in most of the examples presented
so far. See “Initializing arrays” on page 17, “As generic arrays” on page 18, and
“As generic indexed arrays” on page 18 for details.
Tuples
Data structures in OPL can be constructed using tuples that cluster together closely
related data. This topic describes how to declare tuples, use keys on tuples, and
initialize tuples.
Declaring tuples
For example, the declaration:
tuple Point {
int x;
int y;
};
Point point[i in 1..3] = <i, i+1>;
declares a tuple Point consisting of two fields xand yof type integer. Once a tuple
type Thas been declared, tuples, arrays of tuples, sets of tuples of type T, tuples of
tuples can be declared, as in:
Point p = <2,3>;
Point point[i in 1..3] = <i, i+1>;
{Point} points = {<1,2>, <2,3>};
Chapter 2. OPL, the modeling language 9
tuple Rectangle {
Point ll;
Point ur;
}
These declarations respectively declare a point, an array of three points, a set of
two points, and a tuple type where the fields are points. The various fields of a
tuple can be accessed in the traditional way by suffixing the tuple name with a dot
and the field name, as in
Point p = <2,3>;
int x = p.x;
which initializes xto the field xof tuple p. Note that the field names are local to
the scope of the tuples.
Note:
Multidimensional arrays are not supported in tuples.
Keys in tuple declarations
As in database systems, tuple structures can be associated with keys. Tuple keys
enable you to access data organized in tuples using a set of unique identifiers. In
the following example, the nurse tuple is declared with the key name of type
string.
Declaring a tuple using a single key (nurses.mod)
tuple nurse {
key string name;
int seniority;
int qualification;
int payRate;
}
Then, supposing Isabelle must not work more than 20 hours a week, just write:
NurseWorkTime[<"Isabelle">]<=20;
leaving out the fields with no keys. This is equivalent to:
NurseWorkTime[<"Isabelle",3,1,16>]<=20;
Using keys in tuple declarations has practical consequences, in particular:
vThe key field can be used as a unique identifier for the tuple, for example the
field name in Declaring a tuple using a single key (nurses.mod). In this example,
it means that there will be no two tuples with the same name in a set of tuples
of the type nurse. If a user inadvertently attempts to add two different tuples
with the same name, OPL raises an error.
vDefining keys enables you to access elements of the tuple set by using only the
value of the key field (name in the nurses example). Slicing is one of the features
that benefit from it: you can slice on the tuple set using only key fields.
You can also declare a tuple using a non singleton set of keys, such as the shift
tuple of the nurses example below.
Declaring a tuple using a set of keys (nurses.mod)
tuple shift {
key string departmentName;
key string day;
10 OPL Language Reference Manual
key int startTime;
key int endTime;
int minRequirement;
int maxRequirement;
}
In this example, a shift is uniquely identified by the department name, the date,
and start and end times, all defined as key fields.
Initializing tuples
You initialize tuples by giving the list of the values of the various fields, as in:
Point p = <2,3>;
which initializes p.x to 2and p.y to 3. See “Initializing tuples” on page 20 for
details.
Limitations on tuples
When using tuples in your models, you should be aware of various limitations.
Data types in tuples
Not all data types are allowed inside tuples. The limitations are given here.
Data types allowed in tuples
vPrimitives (int, float, string)
vTuples (also known as subtuples)
vArrays with primitive items (not string), that is: integer or float arrays
vSets with primitive items, that is: integer, float or string sets
Note:
Arrays and sets in tuples are not compared by content. If collections inside a tuple
are modified, duplicates are not detected.
Data types not allowed in tuples
vSets of tuples (instances of IloTupleSet)
vArrays of strings, tuples, and tuple sets
vMultidimensional arrays
Tuple indices and tuple patterns
You cannot mix tuple indexes and patterns within the declaration and the use of
decision expressions. For example, these code lines raise the following error
message Data not consistent for "xxx": can not mix pattern and index
between declaration of dexpr and instantiation.
Do not mix tuple indices and tuple patterns in dexpr
dexpr float y[i in t] = ...;
subject to {
forall(<a,b,c> in t) y[<a,b,c>]==...; };
dexpr float y[<a,b,c> in t] = ...;
subject to {
forall(i in t) y[i]==...;
};
Chapter 2. OPL, the modeling language 11
Performance and memory consumption
If you choose to label constraints in large models, use tuple indices instead of tuple
patterns to avoid increasing the performance and memory cost. See “Constraint
labels” on page 53.
Sets
Gives a definition of sets, a list of the operations allowed on sets, and a few words
on their initialization.
Definition
Sets are non-indexed collections of elements without duplicates.
OPL supports sets of arbitrary types to model data in applications. If Tis a type,
then {T}, or alternatively setof(T), denotes the type “set of T”. For example, the
declaration:
{int} setInt = ...;
setof(Precedence) precedences = ...;
declares a set of integers and a set of precedences.
Sets may be ordered, sorted, or reversed. By default, sets are ordered, which means
that:
vTheir elements are considered in the order in which they have been created.
vFunctions and operations applied to ordered sets preserve the order.
See “Sorted and ordered sets” on page 13 for details.
Operations on sets
The following operations are allowed on sets. See OPL functions in Language Quick
Reference for more information about functions. For the functions on sets, the index
starts at 0.
Table 2. Operations allowed on sets
Operations Syntax
union, inter, diff, symdiff set = set1 function set2
first, last elt = function(set)
next, prev elt = function(set,elt,int)
nextc, prevc elt = function(set,elt,int)
item elt = function(set,int)
ord int = function(set,elt)
Initializing sets
A set can be initialized in various ways. The simplest way is by listing its values
explicitly. For example:
tuple Precedence {
int before;
int after;
}
{Precedence} precedences = {<1,2>, <1,3>, <3,4>};
12 OPL Language Reference Manual
See “Initializing sets” on page 21 for details.
Methods for sets
For advanced users:
The methods for sets can be found in the class IloDiscreteDataCollection,inthe
ILOG Script Reference Manual.
Sorted and ordered sets
Shows the effect of the sorted property on simple sets, tuple sets, and sets used in
piecewise linear functions.
Definitions
vAn ordered set is a set whose elements are arranged in the order in which they
were added to the set. Note that this is how sets are created by default. For
example:
{int} S1 = {3,2,5};
and
ordered {int} S1 = {3,2,5};
are equivalent.
vA sorted set is a set in which elements are arranged in their natural, ascending
(or descending) order. For strings, the natural order is the lexicographic order.
The natural order also depends on the system locale. For example:
sorted {int} sortedS = {3,2,5};
and
ordered {int} orderedS = {2,3,5};
are equivalent, and iterating over sortedS or orderedS will have the same
behavior.
To specify the descending order, you add the keyword reversed.
Simple sets
The following code sample enables you to see the difference between the union of
ordered sets and the union of sorted sets.
{int} s1 = {3,5,1};
{int} s2 = {4,2};
{int} orderedS = s1 union s2;
sorted {int} sortedS = s1 union s2;
execute{
writeln("ordered union = ", orderedS);
writeln("sorted union = ", sortedS);
}
The statement
{int} orderedS = s1 union s2;
returns
ordered union = {35142}
while the statement
sorted {int} sortedS = s1 union s2;
returns
Chapter 2. OPL, the modeling language 13
sorted union = {12345}
Sorted tuple sets
When a tuple set does not use keys, the entire tuple, except set and array fields, is
taken into account for sorting. For tuple sets with keys, sorting takes place on all
keys in their order of declaration. In other words, it is not possible to sort a tuple
set on one (or more) given column(s) only.
The code extract below declares a team of people who are defined by their first
name, last name, and nickname, then prints the list of team members first in the
creation order, then in alphabetical order.
tuple person {
string firstname;
string lastname;
string nickname;
}
tuple personKeys {
key string firstname;
key string lastname;
string nickname;
}
{person} devTeam = {
<"David", "Atkinson", "Dave">,
<"David", "Doe", "Skinner">,
<"Gregory", "Simons", "Greg">,
<"David", "Smith", "Lewis">,
<"Kevin", "Morgan", "Kev">,
<"Gregory", "McNamara ", "Mac">
};
sorted {personKeys} sortedDevTeam = {<i,j,k> | <i,j,k> in devTeam};
execute{
writeln(devTeam);
writeln(sortedDevTeam);
}
The person tuple uses no keys.
tuple person {
string firstname;
string lastname;
string nickname;
}
The personKeys tuple uses keys for the first and last names, not for the nickname.
tuple personKeys {
key string firstname;
key string lastname;
string nickname;
}
The data shows that the team includes three people whose first name is David,
two people whose first name is Gregory, and one person whose first name is
Kevin.
As a consequence, the statement
sorted {personKeys} sortedDevTeam = {<i,j,k> | <i,j,k> in devTeam};
lists the David tuples before the Gregory tuples, which themselves appear before
the Kevin tuple. Within the David tuples, "David" "Doe" "Skinner" comes before
"David" "Smith" "Lewis" because a second sorting also takes place on the second
14 OPL Language Reference Manual
field with the key lastname. In contrast, since there is no person with the same first
name and last name, no sort is ever done on the last field nickname.
The output of sortedDevTeam is displayed in the CPLEX Studio IDE as:
{<"David" "Atkinson" "Dave"> <"David" "Doe" "Skinner">
<"David" "Smith" "Lewis"> <"Gregory" "McNamara " "Mac">
<"Gregory" "Simons" "Greg"> <"Kevin" "Morgan" "Kev">}
Sorted sets in piecewise linear functions
In piecewise linear functions, breakpoints must be strictly increasing. However, in
most cases, the data supplied by a database or a .dat file is not sorted in an
increasing numeric or lexicographic order. As a consequence, you have to add
complex and verbose scripting statements to sort the data.
To avoid these extra code lines, the sorted property of sets enables you to sort data
by specifying a single keyword, as shown in the code extract below. Writing
piecewise linear functions becomes easier, as one code line is sufficient instead of
several dozens.
tuple Cost{
key int BreakPoint;
float Slope;
}
sorted {Cost} sS = { <1, 1.5>, <0, 2.5>, <3, 4.5>, <2, 4.5>};
float lastSlope = 3.5;
dvar float+ x;
minimize piecewise(t in sS)
{t.Slope -> t.BreakPoint; lastSlope} x;
See also “Piecewise-linear functions” on page 43.
Other related topics:
“Data sources”
Describes data and database initialization, spreadsheet input/output, data
consistency, and preprocessing.
Describes data initialization.
Introduction to scripting
Defines IBM ILOG Script as a scripting language and describes the situations in
which it is used: preprocessing, postprocessing, and flow control. Also provides
programming tips and warns you of pitfalls to avoid.
Describes how to set declarations.
Data sources
Describes data and database initialization, spreadsheet input/output, data
consistency, and preprocessing.
Data initialization
Defines internal versus external initialization, describes how to initialize arrays,
tuples, and sets, and discusses memory allocation aspects of data initialization.
Internal and external initialization
Defines these two kinds of data initialization.
Chapter 2. OPL, the modeling language 15
In OPL, you can initialize data internally or externally. Your choice affects memory
allocation. See “Initialization and memory allocation” on page 23 for details.
Internally
This initialization mode consists in initializing the data in the model file at the
same time as it is declared. Inline initializations can contain expressions to initialize
data items, such as
int a[1..5] = [b+1, b+2, b+3, b+4, b+5];
Note:
If you choose to initialize data within a model file, you will get an error message if
you try to access it later by means of a scripting statement such as:
myData.myArray_inMod[1] = 2;
Externally
This initialization mode consists in specifying initialization subsequently as an OPL
statement in a separate .dat file (see OPL Syntax in the Language Quick Reference).
This includes reading from a database, as explained in “Database initialization” on
page 24, or from a spreadsheet as explained in “Spreadsheet Input/Output” on
page 29.
You declare external data using the ellipsis syntax. However, data initialization
instructions cannot contain expressions, since they are intended to specify data.
Data initialization instructions make it possible to specify sets of tuples in very
compact ways. Consider these types in a .mod file:
{string} Product ={"flour", "wheat", "sugar"};
{string} City ={"Providence", "Boston", "Mansfield"};
tuple Ship {
string orig;
string dest;
string p;
}
{Ship} shipData = ...;
and assume that the set of shipments is initialized externally in a separate .dat file
like this:
shipData =
{
<"Providence", "Boston", "wheat">
<"Providence", "Boston", "flour">
<"Providence", "Boston", "sugar">
<"Providence", "Boston", "wheat">
<"Providence", "Mansfield", "wheat">
<"Providence", "Mansfield", "flour">
<"Boston", "Providence", "sugar">
<"Boston", "Providence", "flour">
};
Note:
In .dat files, the separating comma is optional. For strings without any special
characters, even the enclosing quotes are optional.
16 OPL Language Reference Manual
Initializing arrays
Describes the various ways in which you can initialize arrays.
You can initialize arrays:
v“Externally”
v“Internally” on page 18
v“In preprocessing instructions” on page 18
v“As generic arrays” on page 18
v“As generic indexed arrays” on page 18
v“From a database” on page 20
Externally
Arrays can be initialized by external data, in which case the declaration has the
form:
int a[1..2] [1..3] = ...;
and the actual initialization is given in a data source, which is a separate .dat file
in IBM ILOG OPL.
vBy listing values
This is how arrays are initialized in most of the examples presented so far.
Multidimensional arrays in OPL are, in fact, arrays of arrays and must be
initialized accordingly. For example, the declaration:
/* .mod file */
int a[1..2][1..3] = ...;
/* .dat file */
a=[
[10, 20, 30],
[40, 50, 60]
];
initializes a two-dimensional array by giving initializations for the
one-dimensional arrays of its first dimension. It is easy to see how to generalize
this initialization to any number of dimensions.
vBy specifying pairs
An array can also be initialized by specifying pairs (index, value), as in the
declaration:
/* .mod file */
int a[Days] = ...;
/* .dat file */
a=#[
“Monday”: 1,
”Tuesday”: 2,
”Wednesday”: 3,
”Thursday”: 4,
”Friday”: 5,
”Saturday”: 6,
”Sunday”: 7
]#;
Note:
1. When the initialization is specified by (index, value) pairs, the delimiters #[
and ]# must be used instead of [and ].
2. The ordering of the pairs can be arbitrary.
These two forms of initialization can be combined arbitrarily, as in:
Chapter 2. OPL, the modeling language 17
/* .mod file */
int a[1..2][1..3] = ...;
/* .dat file */
a=#[
2: [40, 50, 60],
1: [10, 20, 30]
]#;
Internally
You can initialize arrays internally (that is, in the .mod file) using the same syntax
as in .dat files. Here, the array items may be expressions that are evaluated during
initialization. The syntax for pairs #[,]# is not available for internal initialization.
In preprocessing instructions
Arrays can also be initialized in the preprocessing instructions, as in:
range R = 1..8;
int a[R];
execute {
for(var i in R) {
a[i]=i+1;
}}
which initializes the array in such a way that a[1]=2,a[2]=3, and so on.
See “Preprocessing of data” on page 34.
As generic arrays
OPL also supports generic arrays, that is, arrays whose items are initialized by an
expression. These generic arrays may significantly simplify the modeling of an
application. The declaration:
int a[i in 1..10] = i+1;
declares an array of 10 elements such that the value of a[i] is i+1. Generic arrays
can of course be multidimensional, as in:
int m[i in 0..10][j in 0..10] = 10*i + j;
which initializes element m[i][j] to 10*i + j. Generic arrays are useful in
performing some simple transformations. For instance, generic arrays can be used
to transpose matrices in a simple way, as in:
int m[Dim1][Dim2] = ...;
int t[j in Dim2][i in Dim1] = m[i][j];
More generally speaking, generic arrays can be used to permute the indices of
arrays in simple ways.
As generic indexed arrays
To have more flexibility when initializing arrays in a generic way, OPL enables you
to control the index value in addition to the item value, as described earlier in As
generic arrays. To illustrate the syntax, the same examples can be expressed as
follows:
int a[1..10]=[i-1:i|iin2..11 ];
int m[0..10][0..10]=[i:[j:10*i+j ] | i,j in 0..10 ];
18 OPL Language Reference Manual
This syntax is close to the syntax used for initializing arrays in .dat files by means
of indices, delimited by #[ and ]#, as explained above in Specifying pairs. Using
this syntax is an efficient means of initializing arrays used to index data.
The oilDB.mod example contains an execute block that performs initialization.
Instead of:
GasType gas[Gasolines];
execute {
for(var g in gasData) {
gas[g.name] = g
}
}
the same can be expressed with the syntax for generic indexed arrays as:
GasType gas[Gasolines] = [ g.name:g|gingasData ];
Likewise, this syntax, from transp4.mod
float Cost[Routes];
execute INITIALIZE {
for( var t in TableRoutes ) {
Cost[Routes.get(t.p,Connections.get(t.o,t.d))] = t.cost;
}
}
is equivalent to
float Cost[Routes] = [ <t.p,<t.o,t.d>>:t.cost|tinTableRoutes ];
Note:
1. It is recommended to use generic arrays or generic indexed arrays whenever
possible, since they make the model more explicit and readable.
2. If an index is met more than once, no warning is issued and the latest value set
for this index is the one kept.
For example:
int n=5;
{int} s= {1,3,4,2,5};
sorted {int} s2=asSet(1..n);;
reversed {int} s3=asSet(1..n);;
int x[1..n]=[maxl(n-i,i):i|iins];
int x2[1..n]=[maxl(n-i,i):i|iins2];
int x3[1..n]=[maxl(n-i,i):i|iins3];
execute
{
writeln(x);
writeln(x2);
writeln(x3);
}
gives out
[00245]
[00345]
[00215]
Chapter 2. OPL, the modeling language 19
From a database
The example below is more efficient because no data is duplicated.
Reading database columns to a tuple array (oilDB2.dat)
Gasolines,Gas from DBRead(db,"SELECT name,name,demand,price,octane,lead FROM GasData");
Oils,Oil from DBRead(db,"SELECT name,name,capacity,price,octane,lead FROM OilData");
You can also write:
Gasolines from DBRead(db,"SELECT name FROM GasData");
Gas from DBRead(db,"SELECT name,demand,price,octane,lead FROM GasData");
Oils from DBRead(db,"SELECT name from OilData");
Oil from DBRead(db,"SELECT name,capacity,price,octane,lead FROM OilData");
Initializing tuples
Describes the two ways of initializing tuples.
You initialize tuples either by giving the list of the values of the various fields (see
“Tuples” on page 9) or by listing the fields and values. For example:
In the .mod file, you write:
tuple point
{
int x;
int y;
}
point p1=...;
point p2=...;
In the .dat file, you write:
p1=#<y:1,x:2>#;
p2=<2,1>;
As with arrays, the delimiters <and >are replaced by #< and ># and the ordering
of the pairs is not important. OPL checks whether all fields are initialized exactly
once.
The type of the fields can be arbitrary and the fields can contain arrays and sets.
Example 1: tuple Rectangle
For example, the following code lines declare a tuple with three fields: the first is
an integer and the other two are arrays of two points.
tuple Rectangle {
int id;
int x[1..2];
int y[1..2];
}
Rectangle r = ...;
execute
{
writeln(r);
}
A specific “rectangle” can be declared in the data file as:
r=<1, [0,10], [0,10]>;
20 OPL Language Reference Manual
Example 2: tuple Precedence
The declaration
tuple Precedence {
string name;
{string} after;
}
defines a tuple in which the first field is a set item and the second field is a set of
values. A possible precedence can be declared as follows:
Precedence p = <a1, {a2, a3, a4, a5}>;
assuming that a1,..,a5 are strings.
You can also initialize tuples internally within the .mod file. If you choose to do so,
you cannot use the named tuple component syntax #<,>#, which is supported in
.dat files but not in .mod files. Components may be expressions and will be
evaluated during initialization.
Initializing sets
You can initialize sets in three different ways.
v“Externally”
v“Internally”
v“As generic sets” on page 22
Externally
As stated in “Initializing sets” on page 12, in the Sets topic, the simplest way to
initialize a set is by listing its values explicitly in the .dat file.
For example, the declaration:
/* .mod file */
tuple Precedence {
int before;
int after;
}
{Precedence} precedences = ...;
/* .dat file */
precedences = {<1,2>, <1,3>, <3,4>};
initializes a set of tuples.
Internally
You can also initialize sets internally (in the .mod file), more precisely by using set
expressions using previously defined sets and operations such as union,
intersection, difference, and symmetric difference. The symmetric difference of two
sets Aand Bis
(A B)\(A B)
described in “Expressions” on page 41.
For example, the declarations:
{int} s1 = {1,2,3};
{int} s2 = {1,4,5};
{int}i=s1inter s2;
Chapter 2. OPL, the modeling language 21
{int} j = {1,4,8,10} inter s2;
{int}u=s1union {5,7,9};
{int}d=s1diff s2;
{int} sd = s1 symdiff {1,4,5};
initialize ito {1},uto {1,2,3,5,7,9},dto {2,3}, and sd to {2,3,4,5}.
It is also possible to initialize a set from a range expression. For example, the
declaration:
{int} s = asSet(1..10)
initializes sto {1,2,..,10}
It is important to point out at this point that sets initialized by ranges are
represented explicitly (unlike ranges). As a consequence, a declaration of the form
{int} s = asSet(1..100000);
creates a set where all the values 1,2, ..., 100000 are explicitly represented, while
the range
range s = 1..100000;
represents only the bounds explicitly.
Note that when writing the assignment s2=s1, you add one element to s1, that
element is also added to s2. If you do not want this, write
s1={i|i in s2}
For example, compare the statements in the table below.
Table 3. Initializing sets in the model file
If you write {int} s1={1,2};
{int} s2=s1;
execute
{
s2.add(3);
writeln(s1);
}
{int} s1={1,2};
{int} s2={i|iins1};
//{int} s2=s1;
execute
{
s2.add(3);
writeln(s1);
}
the result is {123} {12}
As generic sets
OPL supports generic sets which have an expressive power similar to relational
database queries. For example, the declaration:
{int}s={i|iin1..10: i mod 3 == 1};
initializes swith the set {1,4,7,10}. A generic set is a conjunction of expressions of
the form
p in S : condition
22 OPL Language Reference Manual
where pis a parameter (or a tuple of parameters), Sis a range or a finite set, and
condition is a Boolean expression. These expressions are also used in forall
statements and aggregate operators and are discussed in detail in “Formal
parameters” on page 64.
The declaration:
{string} Resources ...;
{string} Tasks ...;
Tasks res[Resources] = ...;
tuple Disjunction {
{string} first;
{string} second;
}
{Disjunction} disj = {<i,j> |
r in Resources, ordered i,j in res[r]
};
is a more interesting example, showing a conjunction of expressions, and is
explained in detail in “Formal parameters” on page 64. Generic sets are often
useful when you transform a data structure (e.g. the data stored in a file) into a
data structure more appropriate for stating the model effectively. Consider, for
example, the declarations:
{string} Nodes ...;
int edges[Nodes][Nodes] = ...;
which describe the edges of a graph in terms of a Boolean adjacency matrix. It may
be important for the model to use a sparse representation of the edges (because,
for instance, edges are used to index an array). The declaration:
tuple Edge {
Nodes o;
Nodes d;
}
{Edge} setEdges = {<o,d> | o,d in Nodes : edges[o][d]==1};
computes this sparse representation using a simple generic set. It is of course
possible to define generic arrays of sets. For example, the declaration:
{int} a[i in 3..4] = {e|ein1..10: e mod i == 0};
initializes a[3] to {3,6,9} and a[4] to {4,8}.
Initialization and memory allocation
Describes how memory is allocated to data initialization.
In OPL, the initialization mode you choose affects memory allocation. Namely,
external initialization from a .dat file, while enabling a more modular design, may
have a significant impact on memory usage.
Internal initialization
Internal data (directly from the model file) is initialized when first used. This is
also called “lazy initialization”. Unused internal data elements are not allocated
any memory. In other words, internal data is “pulled” from OPL as needed.
Example of lazy initialization
int a=2;
int b=2;
int a2=2*a;
Chapter 2. OPL, the modeling language 23
int b2=2*b;
execute
{
a2;
a++;
b++;
writeln(a2);
writeln(b2);
}
assert a2==4;
assert b2==6;
External initialization
In contrast, data from a data file is initialized while the .dat file is parsed and is
allocated memory whether it is used by the model or not. In other words, external
data is “pushed” to OPL.
See also “Pitfall of lazy initialization of data” on page 36.
Database initialization
Describes how to connect to one or several relational databases, how to read from
such databases using traditional SQL queries, and to write the results back to the
connected database.
The oil database example
Explains database initialization in the context of an oil database.
The syntax for databases is valid only for data files, with the extension .dat, not
for model files with the extension .mod. This section uses the oilDB example to
demonstrate operations with a Microsoft Access database. You can find this
example in
<Install_dir>/opl/examples/opl/oil
where <Install_dir> is your installation directory.
Working with databases (oilDB.dat)
DBConnection db("access","oilDB.mdb");
Gasolines from DBRead(db,"SELECT name FROM GasData");
Oils from DBRead(db,"SELECT name FROM OilData");
GasData from DBRead(db,"SELECT * FROM GasData");
OilData from DBRead(db,"SELECT * FROM OilData");
MaxProduction = 14000;
ProdCost = 4;
DBExecute(db,"drop table Result");
DBExecute(db,"create table Result(oil varchar(10), gas varchar(10), blend real, a real)");
Result to DBUpdate(db,"INSERT INTO Result(oil,gas,blend,a) VALUES(?,?,?,?)");
Connection to a database
Shows how to connect OPL to a database.
In OPL, database operations all refer to a database connection. Here are two
examples from the oilDB example for declaring connections. See “Database
connection strings” on page 25 for more connection strings.
DBConnection db("odbc","oilDB/user/passwd");
24 OPL Language Reference Manual
and
DBConnection db("access","oilDB.mdb");
The first example uses the ODBC data source oilDB declared by the system to
connect to the database.
The connection db should be viewed as a handle on the database.
Note:
1. The user and passwd parameters are optional: you can connect to oilDB//
without a user name and password.
2. It is possible to connect to several databases within the same model.
Database connection strings
Provides the syntax of strings used to connect to the databases supported by OPL.
Supported databases in the Working Environment document provides a list of the
databases to which you can connect your OPL model in order to read and write
data.
The table below gives the syntax of the string you must use to connect to each of
the supported databases.
Table 4. Database connection strings
Database Name Connection String Driver
DB2 DBConnection db
("db2","username/password/
database");
(The client configuration will
find the server.)
DB2
Oracle 10g
Oracle 11g
DBConnection
db("oracle10","userName/
password@dbInstance")
DBConnection
db("oracle11","userName/
password@dbInstance");
Oracle
MS SQL Server DBConnection
db("oledb","userName/
password/database/
dbServer");
OLE DB
MS Access DBConnection
db("access","<filename>.mdb");
or
DBConnection
db("odbc","dataSourceName/
userName/password");
ODBC
Other databases not listed
above
DBConnection
db("odbc","dataSourceName/
userName/password");
ODBC
Chapter 2. OPL, the modeling language 25
See The oil database example in IDE Tutorials for details on how to customize the
example for a connection to a different database.
Reading from a database
Explains the process of reading data from a database in OPL.
In OPL, database relations can be read into sets or arrays. For instance, these
instructions from the model file:
tuple gasType {
string name;
float demand;
float price;
float octane;
float lead;
}
tuple oilType {
string name;
float capacity;
float price;
float octane;
float lead;
}
And these instructions from the data file:
GasData from DBRead(db,"SELECT * FROM GasData");
OilData from DBRead(db,"SELECT * FROM OilData");
Together illustrate how to initialize a set of tuples from the relation OilData in the
database db. In this example, the DBRead instruction inserts an element into the set
for each tuple of the relations.
Important conventions adopted by OPL:
1. If read into a set, the resulting set must be a set of integers, floats, or strings, or
a set of tuples whose elements are integers, floats, or strings.
2. If read into an array, the resulting array must be an array of integers, floats, or
strings, or an array of tuples whose elements are integers, floats, or strings.
3. In the case of tuples, the columns of the SQL query result are mapped by
position to the field of the OPL tuples. For instance, in the above query, the
column name has been mapped to the field name and so on.
4. When initializing an array with a DBRead statement, the indexing set and array
cells are initialized at the same time.
Note:
OPL does not parse the query; it simply sends the string to the database system
that has full responsibility for handling it. As a consequence, the syntax and the
semantics of these queries are outside the scope of this book and users should
consult the appropriate database manual for more information.
It is also possible to implement parameterized queries in OPL, for example:
Oils from DBRead(db,"SELECT name FROM OilData WHERE quality>?")(oilQuality);
where oilQuality is any scalar OPL data element already initialized and whose
type is expected in the SQL query. In this case, oilQuality should be a numeric
type, for example an integer.
26 OPL Language Reference Manual
Note:
Despite standardization, Oracle does not support the question mark as a variable
identifier. Use ':'<parameter number> instead. Examples are ':1', ':arg', etc.
SQL encryption:
In OPL 3
Because all database instructions were in the model file, the SQL statements were
encrypted as well when the model was compiled.
In OPL 4 and later
To do the same in OPL 4.x (where you write database instructions in data files),
you can define literal strings inside the model file (which will be compiled) and
use them in the data file, like this:
In the .mod file:
string connectionString = "scott/tiger@TEST";
string myQuery = "select id from table";
{int} setOfInt = ...;
dvar int X in 1..5;
minimize X;
subject to {
forall (i in setOfInt)
X>=i;
};
In the .dat file:
DBconnection db("oracle9", connectionString);
setOfInt from DBread (db, myQuery);
Writing to a database
Explains the process of writing to a database from OPL.
Publishing results to a database is similar to parameterized data initialization. Here
is an example extracted from the oil code sample:
All database publishing requests are carried out during postprocessing, if a
solution is available. Such requests are processed in the order declared in the .dat
file(s). If your RDMBS supports transactions, every single publishing request is
sent within its own transaction.
Adding rows
1. Write in the model file:
tuple result {
string oil;
string gas;
float blend;
float a;
}
{result} Result =
{ <o,g,Blend[o][g],a[g]>|oinOils, g in Gasolines };
2. Write in the data file:
Chapter 2. OPL, the modeling language 27
DBExecute(db,"drop table Result");
DBExecute(db,"create table Result(oil varchar(10), gas varchar(10), blend real, a real)");
Result to DBUpdate(db,"INSERT INTO Result(oil,gas,blend,a) VALUES(?,?,?,?)");
In this example, you use:
vaDBExecute statement to send SQL DDL (data definition language) instructions
to the Relational Database Management Server (RDBMS)
vaDBUpdate statement to modify the data (see “Updating existing rows”).
More generally, the keyword DBExecute enables you to carry out administration
tasks on data tables, whereas the keyword DBUpdate modifies the contents of data
tables.
The OPL result publisher will iterate on the items in the set result and bind the
component values to the SQL statement parameters in the declared order.
Note:
OPL supports the same element types for reading as for database publishing.
Updating existing rows
To update existing rows in a database instead of adding new ones, use an SQL
update statement.
For example, to multiply by 2 the blends for Super:
1. Add the following lines in the .mod file:
tuple Result2 {
float blend;
float a;
string oil;
string gas;
}
{Result2} result2 = { <2*blend[o]["Super"],a["Super"],o,"Super">|oin
Oils};
2. Write an SQL update statement like this:
result2 to DBUpdate(db,
"UPDATE Result SET blend=? , a=? WHERE oil=? AND gas=?");
See also Getting the data elements from an IloOplModel instance in the Language
User’s Manual for details about data publishers and postprocessing.
Deleting elements
It is also possible to delete elements from a database. For instance, the instructions
/* .mod file */
{string} NamesToDelete = ...;
/* .dat file */
NamesToDelete to DBUpdate(db,"delete from PEOPLE where NAME = ?");
delete from the relation table PEOPLE all the tuples whose names are in
NamesToDelete.
Note:
The syntax of the actual queries may differ from one database system to another.
28 OPL Language Reference Manual
Spreadsheet Input/Output
Describes how to connect an MS Excel spreadsheet, read from it, and write the
results to the connected spreadsheet.
The oilsheet example
Explains spreadsheet input and output in the context of an oil spreadsheet.
This section uses the oilSheet example to demonstrate operations with an MS
Excel spreadsheet. You can find this example in
<Install_dir>/opl/examples/opl/oil
where <Install_dir> is your installation directory.
Using spreadsheets through ODBC
If you access spreadsheet data through an ODBC connection using a JDBC-ODBC
client, the ODBC driver returns NULL if the data is not of the right type instead of
reporting a specific data type error. See http://support.microsoft.com/kb/194124/
EN-US/ for details.
Connection to a spreadsheet
Explains how to connect OPL to a spreadsheet.
The spreadsheet operations in OPL all refer to a spreadsheet connection. The
instruction
/* .dat file */
SheetConnection sheet("transport.xls");
establishes a connection sheet to a spreadsheet named transport.xls. The
connection sheet should be viewed as a handle on the spreadsheet. It is possible in
OPL to connect to several spreadsheets within the same model.
The instruction SheetConnection takes only one parameter and you do not need to
specify the full path to the spreadsheet name. Relative paths are resolved using the
current directory of .dat files.
Note that, in this section, we often use the word “spreadsheet” for “spreadsheet
connection”.
Reading from a spreadsheet
Explains how to read from a spreadsheet from within OPL.
In OPL, spreadsheet ranges can be read into one- or two-dimensional arrays or
sets. For instance, the instructions:
/* .mod file */
{string} Gasolines = ...;
tuple GasType {
float demand;
float price;
float octane;
float lead;
}
GasType gas[Gasolines] = ...;
/* .dat file */
Chapter 2. OPL, the modeling language 29
SheetConnection sheet("oilSheet.xls");
Gasolines from SheetRead(sheet,"’gas data’!A2:A4");
Gas from SheetRead(sheet,"’gas data’!B2:E4");
What data can be read from an Excel spreadsheet?
OPL opens a spreadsheet in read-only mode to read data from it.
The types of data elements supported are:
vsets with integers, floats, strings, or tuples;
vscalar integers, floats, or strings;
varrays with integers, floats, one- or two-dimensional strings, or one-dimensional
tuples;
vone- or two-dimensional arrays of simple types: for such arrays, the data must
be formatted, that is, it must have the same width/length as the array to be
filled. OPL automatically determines whether the data must be read line by line
or column by column. When facing a square zone (a two-dimensional array with
[x][x] as dimensions), the engine reads the data line by line.
Note:
vOnly tuples with integer, float, and string components are supported.
vIf the sheet name contains a space, the name should be surrounded by single
quotes in the SheetRead instruction. For example:
Oil from SheetRead(sheet,"’oil data’!B2:E4");
vOPL does not support the R1C1 reference style to specify the range when
reading data from an Excel spreadsheet.
Accessing named ranges in Excel
IBM ILOG OPL supports the convention of names, which are a word or string of
characters used to represent a cell, range of cells, formula, or constant value, and
that can be used in other formulas.
Thus you can use easy-to-understand names, such as Nutrients, to refer to
hard-to-understand ranges, such as B4:J15 or IncreasedProtein to refer to a
constraint. You can then substitute these names in formulas for the range of cells
or constraint.
Excel names, or named ranges, can be accessed using the SheetRead command,
using the following syntax:
SheetConnection sheetData("C:\\ILOG_Files\\myExcelFile.xls");
prods from SheetRead(sheetData,"Product");
Note the double separator \\ in the SheetConnection command.
The SheetRead command is normal, and in this example the Excel name Product
replaces the normal syntax of, say, C13:O72.
To create named ranges in Excel 2007:
1. Highlight the range of cells you want to name, then click the Name box at the
left end of the Formula Bar.
2. Type the name you want to assign to this range and press Enter.
3. Save the spreadsheet file.
30 OPL Language Reference Manual
Note: For a list of supported versions of Microsoft Excel, please see the detailed
system requirements http://www-01.ibm.com/support/
docview.wss?uid=swg27019100.
Additional information on named ranges
vExcel automatically updates (expands) a named range when a row is added
somewhere within the range. However, one must careful adding rows at the end
of a range as the range does not get automatically updated in that case. It would
have to be updated manually.
vOPL allows blank rows in a named range. If you are reading a set of strings, it
will consider the blank cells as having the value 0. If you are reading a set of
strings, then it inserts an empty string "" into the set. For example:
s2 = {"Monday" "" "Wednesday" "Thursday" "Friday"}
This behavior is the same when you don't use named range but instead use
explicit ranges like C1:C5, where C2 is empty.
vWith the Excel VBA one can name the first (top left) cell of a named range and
access the whole range. OPL does not support this.
vWhen using sheetWrite to write to named ranges, the size of the range does not
have to match the size of the data you are writing to Excel. If the set is smaller,
then only the top most cells will be filled.
If you try to write more data than the range can accommodate, you receive the
error message: "Exception from IBM ILOG Concert: excel: range is not wide
enough to write the set".
In this sense, named ranges behave in exactly the same way as "regular" ranges.
Format of the Excel data
Here we must differentiate between simple types and tuples:
vSets of simple types: The engine reads data from left to right and top to bottom.
Data can therefore be read either horizontally, vertically, or from a rectangular
zone.
vSets of tuples: The data has to be formatted because the tuple schema has an
arity. As in databases and manual tables, the data format is “fixed width,
variable length”. Therefore, tuple sets are read only line by line in Excel: this is
the same representation as in pure data files.
Writing to a spreadsheet
Explains how to write to a spreadsheet from within OPL.
This section uses extracts from the oilSheet.dat data file.
Publishing results to a spreadsheet can be performed using instructions such as:
a to SheetWrite(sheet,"RESULT!A2:A4");
blend to SheetWrite(sheet,"RESULT!B2:D4");
OPL then opens spreadsheets in read-write mode. This action may fail if another
process is already using the .xls file.
The types of data elements supported for writing are just the same as for reading.
Cells in Microsoft Excel spreadsheets are filled from left to right and from top to
bottom.
Excel names (or named ranges) can be accessed using the SheetWrite command,
using the following syntax:
Chapter 2. OPL, the modeling language 31
SheetConnection sheetData("C:\\ILOG_Files\\myExcelFile.xls");
prods to SheetWrite(sheetData,"Product");
Note the double separator \\ in the SheetConnection command. The SheetWrite
command is normal, and, in this example, the Excel name Product replaces the
normal syntax of, say, C13:O72.
Note:
vOPL does not support the R1C1 reference style to specify the range when
writing data to an Excel spreadsheet.
For more information on named ranges, see “Accessing named ranges in Excel” on
page 30.
Data consistency
Defines the purpose of data consistency and describes data membership and
assertions as ways to ensure consistency.
Purpose
Provides an overview of data consistency issues in OPL.
For an optimization problem to give relevant solutions, it is fundamental that you
provide good quality data to your projects. In particular, it may be interesting to
check that the data is consistent. If the project data is not consistent, the solving
engine may find a wrong solution, or no solution, and you may think that the
model is erroneous and therefore waste time trying to improve it.
OPL offers several ways to check the consistency of the data used by your projects.
In particular:
v“Data membership consistency”: use the keyword with to ensure that cells of a
given tuple in a tuple set correctly belong to a given set of possible values.
Note:
You can also use the keyword key for data consistency when declaring tuples.
See “Keys in tuple declarations” on page 10.
v“Assertions” on page 34: use the keyword assert to ensure that some assertions
on the data are respected.
Data membership consistency
Explains the use of the with keyword to ensure data consistency.
The keyword with enables you to indicate that a given element of a tuple must be
contained in a given set. If you use it, OPL checks the consistency of the tuple set
at run time when initializing the set. The syntax is:
{tupletype} tupleset with cell1 in set1, cell2 in set = ...;
Let's take an example. You have a set of arcs between nodes. Nodes are defined by
a tuple set of tuples consisting of an origin node and a destination node. The with
syntax enables you to ensure that the origin and destination nodes do belong to a
given set of nodes. Compare “Data found inconsistent (keyword with)” on page 33
and “Data found consistent (keyword with)” on page 33:
32 OPL Language Reference Manual
Data found inconsistent (keyword with)
{int} nodes = {1, 5, 7};
tuple arc {
int origin;
int destination;
}
{arc} arcs2 with origin in nodes, destination in nodes =
{<1,4>, <5,7>};
execute {
writeln(arcs2);
};
Data found consistent (keyword with)
{int} nodes = {1, 5, 7};
tuple arc {
int origin;
int destination;
}
{arc} arcs1 with origin in nodes, destination in nodes =
{<1,5>, <5,7>};
execute {
writeln(arcs1);
};
If you write “Data found inconsistent (keyword with),” an error will be raised
when the set arcs2 is initialized because the with syntax will detect that the
statement
(int) nodes = (1, 5, 7);
is not consistent with the statement
with origin in nodes, destination in nodes =
{<1,4>, <5,7>}
If you write “Data found consistent (keyword with),” the initialization of the set
arcs1 will work properly because the with syntax will find that the statement
(int) nodes = (1, 5, 7);
is consistent with the statement
with origin in nodes, destination in nodes =
{<1,5>, <5,7>}
Initializing tuple sets referring to other sets
To initialize tuple sets that refer to other sets with keys for data consistency, you
must use initialization expressions that provide only those key values, as shown in
“Initializing tuple sets referring to other sets.” This is true if you initialize those
tuple sets as internal data or as external data in .dat files, databases, or
spreadsheets.
Initializing tuple sets referring to other sets
tuple node
{
key int node_id;
string city;
Chapter 2. OPL, the modeling language 33
string country;
}
{node} nodes = {<1,"Paris","France">,<5,"Madrid","Spain">, <7,"New York","USA">};
tuple arc {
node origin;
node destination;
}
{arc} arcs1 with origin in nodes, destination in nodes=...;
execute {
writeln(arcs1);
};
Assertions
Explains the use of assertions with regard to data consistency.
OPL provides assertions to verify the consistency of the model data. This
functionality enables you to avoid wrong results due to incorrect input data. In
their simplest form, assertions are simply Boolean expressions that must be true;
otherwise, they raise an execution error. For instance, it is common in some
transportation problems to require that the demand matches the supply. The
declaration
int demand[Customers] = ...;
int supply[Suppliers] = ...;
assert sum(s in Suppliers) supply[s] == sum(c in Customers) demand[c];
makes sure that the total supply by the suppliers meets the total demand from the
customers. This assertion can be generalized to the case of multiple products, as in
int demand[Customers] [Products] = ...;
int supply[Suppliers] [Products] = ...;
assert
forall(p in Products)
sum(s in Suppliers) supply[s][p] == sum(c in Customers) demand[c][p];
This assertion verifies that the total supply meets the total demand for each
product. The use of assertions is highly recommended, since they make it possible
to detect errors in the data input early, avoiding tedious inspection of the model
data and results.
Assertions can be labeled. See “Labeled assert statements” on page 55.
Note:
In OPL, elements are initialized when they are first used. This is called 'lazy
initialization' (see “Data initialization” on page 15).
However, the execution of an assert statement triggers an evaluation of the
elements involved in the assertion. So, if you use an OPL element for the first time
in a skipped assert expression, the lazy initialization will be triggered later.
Skipping or removing assert statements may impact the order in which OPL
elements are initialized and thus change the behavior of a program.
Preprocessing of data
Provides an overview of preprocessing operations in OPL.
34 OPL Language Reference Manual
You can preprocess data before the optimization model is created by using IBM
ILOG Script/JavaScript syntax encapsulated in execute blocks.
OPL provides script integration with the modeling language. All declared model
elements are available for scripting via their name.
The functionality available for an element depends on its type. All elements can be
read, but modifications are possible only for primitive types (int,float,string)
and primitive items of arrays and tuples. See the intro to the Reference Manual of
IBM ILOG Script Extensions for OPL about these limitations.
You can change the domain boundaries for decision variables, as well as their
priority, in the preprocessing phase.
You can also use preprocessing to change CPLEX or CP Optimizer parameter
settings (see Changing option values in the Language User’s Manual).
Elements of a range or constraint type are immutable.
Example:
int n = ...;
range R = 1..n;
int A[R] = ...
execute {
for(r in R) {
if ( A[r]<0 ) {
A[r] = 0;
}
}
}
Initialization and processing order
Preprocessing items are processed according to their category, not in absolute
declaration order.
Within categories, the order is:
1. External data, in the order in which the data sources are added to the OPL
model
2. All execute blocks and assert statements, in declaration order
Internal data is initialized as encountered during external data initialization, in
execute blocks and in assert statements. The following examples use internal data.
If you write:
{int} s1={1,2};
{int} s2={i|iins1};
execute
{
writeln(s2);
s1.add(3);
writeln(s1,s2);
}
the result is:
{1 2}
{123}{12}
Chapter 2. OPL, the modeling language 35
whereas if you write:
{int} s1={1,2};
{int} s2={i|iins1};
execute
{
//writeln(s2);
s1.add(3);
writeln(s1,s2);
}
the result is:
{123}{123}
The order in which you use the internal data is important. The order in which you
use external data is not important because it is all created, or initiated, at the
beginning of the run.
See also “Pitfall of lazy initialization of data.”
Use the profiler feature to inspect the instantiation sequence of your model. See
Profiling the execution of a model in IDE Tutorials.
See Chapter 3, “IBM ILOG Script for OPL,” on page 91 for details on the scripting
language and its extensions for OPL.
Lazy initialization
It is important to be aware from OPL 5.2 onwards that during processing, declared
data elements are initialized on demand when referenced for the first time. See
Data preprocessing in Migration from OPL 3.x (CP projects) for migration aspects.
Pitfall of lazy initialization of data
In OPL, external data is always initialized before internal data.
If you are not aware of the lazy initialization of data in OPL, you may be surprised
by unexpected results.
OPL is a declarative language. The order of elements as they are written in the
source file is less important than with imperative languages. We initialize OPL data
elements when they are first used, which may result in an initialization order
different from the declaration order in a .mod file. External data (with =... syntax)
is always initialized before internal data. Scripting blocks are executed after
external data.
Example
In the following example, the array arrayInitExternally is initialized with three
elements in the .dat file, but, because set1 has four elements, the result shows that
arrayInitExternally has one more element (initialized by default with zero).
string arrayInitByScript[0..3];
execute initializeArray {
for(var i=0; i<=3; i++) {
if(i >= 1) arrayInitByScript[i] = "A";
else arrayInitByScript[i] = "B";
}
}
36 OPL Language Reference Manual
// set1 is later used to index an array initialized externally.
// This external initialization of arrayInitExternally alters the content of set1
// because external initialization is done prior to script execution.
// So instead of having set1={1 2 3}, set1 has an additional element "0"
// Before the script is executed, arrayInitByScript contains four empty elemens ["" "" "" ""]
{int} set1 = union(s in 0..3 : arrayInitByScript[s] != "B") {s};
{int} set2 = union(s in 0..3 : arrayInitByScript[s] != "B") {s};
int arrayInitExternally[set1] = ...; // [0, 0, 0]; //
execute {
writeln(arrayInitByScript);
writeln(set1);
writeln(set2);
writeln(arrayInitExternally);
}
This produces the output:
["B" "A" "A" "A"]
{0123}
{123}
[0000]
Decision types
Variables in an OPL application are decision variables (dvar). OPL also supports
decision expressions, that is, expressions that enable you to reuse decision variables
(dexpr). A specific syntax is available in OPL to dynamically collect elements into
arrays.
Decision variables
Shows how to declare and use decision variables in the OPL language.
A decision variable is an unknown in an optimization problem. It has a domain,
which is a compact representation of the set of all possible values for the variable.
Decision variable types are references to objects whose exact nature depends on the
underlying optimizer of a model. A decision variable can be instantiated only in
the context of a given model instance.
The purpose of an OPL model is to find values for the decision variables such that
all constraints are satisfied or, in optimization problems, to find values for the
variables that satisfy all constraints and optimize a specific objective function.
Variables in OPL are thus essentially decision variables and differ fundamentally
from variables in programming languages such as Java, and ILOG Script.
Note:
OPL decision variables are noted with the dvar keyword while the keyword var
denotes ILOG Script variables.
A decision variable declaration in OPL specifies the type and set of possible values
for the variable. Once again, decision variables can be of different types (integer,
float) and it is possible to define multidimensional arrays of decision variables. The
declaration
dvar int transp[Orig][Dest] in 0..100;
declares a two-dimensional array of integer variables. The decision variables are
constrained to take their values in the range 0..100 ; i.e., any solution to the model
Chapter 2. OPL, the modeling language 37
containing this declaration must assign values between 0 and 100 to these
variables. Note that all integer variables need a finite range in OPL. Arrays of
decision variables can be constructed using the same index sets as arrays of data.
In particular, it is also possible, and desirable for larger problems, to index arrays
of decision variables by finite sets. For example, the excerpt:
tuple Route {
City orig;
City dest
}
{Route} routes = ...;
dvar int transp[routes] in 0..100;
declares an array of decision variables transp that is indexed by the finite set of
tuples routes. Genericity can be used to initialize the domain of the variables. For
example, the excerpt:
tuple Route {
City orig;
City dest;
}
{Route} routes = ...;
int capacity[routes] = ...;
dvar int transp[r in routes] in 0..capacity[r];
declares an array of decision variables indexed by the finite set routes such that
variable transp[r] ranges over 0..capacity[r]. The array capacity is also indexed
by the finite set routes. Note that decision variables can be declared to range over
a user-defined range. For example, the excerpt:
range Capacity = 0..limitCapacity;
dvar int transp[Orig][Dest] in Capacity;
declares an array of integer variables ranging over Capacity.
Decision variables can of course be declared individually, as in:
dvar int averageDelay in 0..maxDelay;
For convenience, OPL proposes the types float+,int+ and boolean to define the
domain of a decision variable. The declarations
dvar int+ x; // non negative integer decision variable
dvar float+ y; // non-negative decision variable
dvar boolean z; // boolean decision variable
are therefore equivalent to
dvar int x in 0..maxint;
dvar float y in 0..infinity;
dvar int z in 0..1;
Decision variables in an array can be assigned item-specific ranges, as in
dvar float transp[o in Orig][d in Dest] in 0..cap[o][d];
which declares a two-dimensional array of float variables, where variable
transp[o][d] ranges over the set 0..cap[o][d].
Expressions of decision variables
Shows how to declare and use decision variable expressions in the OPL language.
38 OPL Language Reference Manual
The keyword dexpr allows you to create reusable decision expressions. Indeed, if
an expression has a particular meaning in terms of the original problem, writing it
as a decision expression (dexpr) makes the model more readable.
For example, the scalableWarehouse.mod example expresses the total fixed costs as
a decision expression:
dexpr int TotalFixedCost = sum( w in Warehouses ) Fixed * Open[w];
dexpr float TotalSupplyCost = sum( w in Warehouses, s in Stores ) SupplyCost[s][w] * Supply[s][w];
This way, the two total cost expressions defined are shown in the OPL IDE along
with their values.
You can also use arrays of decision expressions. For example:
dexpr int slack[i in r] = x[i] - y[i];
This array is handled efficiently as only the “definition” is kept. Not all the
expressions for each value of the indices are created. As a consequence, you cannot
change the definition of the dexpr for a particular element of the array.
Using decision expressions is particularly useful and recommended if you plan to
write objectives to be used with ODM Enterprise. Please refer to the ODM
Enterprise documentation.
Dynamic collection of elements into arrays
Discusses the “all” syntax, explicit arrays, appending arrays, and dynamic
initialization of decision variable arrays.
Introduction
Provides an overview of how elements are collected into arrays in OPL.
Some expressions (such as count) and constraints (such as allDifferent) need
arrays of variables or constants to be created. In some models, these expressions or
constraints can be used in an aggregate statement (for example, in a forall
statement) and the exact content of the arrays depends on the iteration.
The all syntax
Shows how to use the all syntax in the OPL language.
Then, the all syntax allows you to dynamically collect some decision variables or
constants into an array. The syntax is similar to sum and forall, it contains a series
of possible generators (an index and a set or a range in which this index is to be
contained), some possible filters (to filter out some of the enumerated
combinations), and a body (here of the form x[i][j]...). The variables or values
in the resulting array follow the logical order of enumerating the index
combinations as defined by the generators.
By default, this dynamic array is indexed from 0to numberOfElements-1. As some
constraints make a particular usage of the index, it may be interesting to define
another indexing schema. For this, it is possible to dynamically define the range of
the resulting array of variables by using the syntax [minindex..maxindex]. Finally,
it is possible to use "*" as maxIndex to indicate that only the minIndex is defined;
the maxIndex will be set accordingly depending on the number of elements.
Here is a complete of usage of the syntax:
Chapter 2. OPL, the modeling language 39
using CP;
intn=5;
range R = 1..n;
dvar int x[R] in R;
subject to {
allDifferent(all(i in R:i%2==1) x[i]);
}
This is a way of defining an array of variables or values, and in all the constraints
and expressions that take arrays. You can use either this syntax or pass directly a
named array. When you pass a named array and indexes make sense in the
constraint, its indexer will be used to index the elements if it has one dimension
only. If it has two dimensions, the indexer cannot be used.
Explicit arrays
Describes explicit arrays in OPL.
Another useful syntax to dynamically create arrays to be used in expressions or
constraints is to explicitly define the array using the []notation and including
any variables or values into it.
Mixes are not allowed. For example, you can write:
forall(i in R)
allDifferent([x[i], y[i], z[i]);
Appending arrays
You can concatenate several arrays using the append function.
For example, if you want to express that all variables from array xand array yare
different, you can use an allDifferent constraint applied to the appended arrays,
as shown in this example.
Appending arrays
using CP;
range R=1..10;
dvar int x[R] in 0..20;
dvar int y[R] in 0..20;
minimize sum(i in R)(x[i]+y[i]);
subject to
{
allDifferent(append(all(i in 1..2) x[i],all(i in 4..6) y[i]));
}
Initialization of decision variable arrays
Shows how to initialize your decision variable arrays in OPL.
The dynamic collection of decision variables allows you to dynamically initialize
an array of decision variables. The variables are then shared between the two
arrays of variables.
Here is an example of what is possible:
intn=5;
range R = 1..n;
dvar int x[i in R] in R;
40 OPL Language Reference Manual
dvar int y1 = x[1];
dvar int y[R] = all[R](i in R) x[i];
dvar int y2[i in R] = x[i];
dvar int y3[R] = [ i:x[i]|iinR];
dvar int y4[R] = [ x[1], x[2], x[3], x[4], x[5] ];
dvar int y5[0..n-1] =
append(all(i in R: i<2) x[i], all(i in R: i>=2) x[i]);
Expressions
Describes data and decision variable identifiers, integer and float expressions,
aggregate operators, piecewise-linear functions (continuous and discontinuous), set
expressions, and Boolean expressions.
Usage of expressions
Describes how to use expressions in OPL.
Expressions are used in fundamentally different ways in OPL:
vto specify items in generic arrays and sets (described in this chapter)
vto filter iterations (see “Formal parameters” on page 64)
vto state constraints over decision variables (see “Constraints” on page 51)
In the first two cases, the expressions do not contain decision variables, since
decision variables have no value at this stage of the computation. These
expressions are said to be ground and they are subject to almost no restrictions.
In the second case, of course, the expressions may contain decision variables.
Boolean expressions containing decision variables are called constraints and are
subject to a number of restrictions; for example, float constraints must be linear,
piecewise-linear, or quadratic.
Data and decision variable identifiers
Describes the use of identifiers within OPL expressions.
Since data and decision variable identifiers are the basic components of
expressions, we will review briefly here how they are used to build expressions. If
ris a tuple with a field capacity of type T, then r.capacity is an expression of
type T.Ifais an n-dimensional array of type T,a[e 1 ]...[e n]is an expression of
type T, provided that eiis well-typed. For instance, the excerpt
int limit[routes] = ...;
dvar int transp[r in routes] in 0..limit[r];
contains an expression limit[r] of type integer. Indices of arrays can be complex
expressions. For instance, the excerpt
int nbFlights = ...;
range Flight = 1..nbFlights;
{string} Employee = ...;
dvar int crew[Flight][Employee] in 0..1;
constraints {
forall(e in Employee)
forall(i in 1..nbFlights - 2)
crew[i][e] + crew[i+1][e] + crew[i+2][e] >= 1;
}
Chapter 2. OPL, the modeling language 41
contains an integer expression crew[i+1][e] whose first index is itself an integer
expression.
Integer and float expressions
Describes the use of constants, data, decision variables, and operators within OPL
expressions.
Integer expressions
Integer expressions are constructed from integer constants, integer data, integer
decision variables, and the traditional integer operators such as +, -, *, div, mod (or
%). The operator div represents the integer division (for example, 8div3==2)
and the operator mod or %represents the integer remainder. OPL also supports the
function abs, which returns the absolute value of its argument, and the built-in
constant maxint, which represents the largest integer representable in OPL.
Note that expressions involving large integers may produce an overflow. In the
following example for int, an overflow is detected.
int a=maxint+2;
float b=infinity+2;
execute
{
writeln(a);
writeln(b);
}
The Problems tab in the IDE displays the error messages.
Most int expressions (such as % or div) are not available for constraints defined in
CPLEX models, but are available for CP models. See also “Constraints available in
constraint programming” on page 61.
Float expressions
Float expressions are constructed from floats, float data and variables, as well as
operators such as +, -, /, *. In addition, OPL contains a float constant infinity to
represent ∞and a variety of float functions, depicted in OPL functions in the
Language Quick Reference.
42 OPL Language Reference Manual
Conditional expressions
Conditional expressions are expressed like this:
(condition)?thenExpr : elseExpr
where condition is a ground condition with no decision variable. If condition is
true, the condition evaluates to thenExpr ; otherwise, it evaluates to elseExpr.
Examples
int value = ...;
int signValue = (value>0)?1:(value<0) ? -1 : 0;
int absValue = (value>=0) ? value : -value;
See the numeric functions in Summary table of OPL functions in the Language
Quick Reference.
Counting expressions
Among integer expressions, there are also some combinatorial expressions. For
example, you can use the count function to count the number of times a particular
value appears in an array of decision variables. You can use such an expression in
modeling constraints only if the modeling constraints are part of a model that is
solved by the CP Optimizer engine (that is, starting with the using CP; statement).
The constraint
count(x, 2) == 3;
states that in the array of variables x, exactly three variables take the value 2.
For more information, see count in the Language Quick Reference.
Aggregate operators
Describes the operators available for computing integer and float summations.
Integer and float expressions can also be constructed using aggregate operators for
computing summations (sum), products (prod), minima (min), and maxima (max)of
a collection of related expressions. For instance, the excerpt
int capacity[Routes] = ...;
int minCap = min(r in Routes) capacity[r];
uses the aggregate operator min to compute the minimum value in array capacity.
The form of the formal parameters in these aggregate operators is very general and
is discussed at length in “Formal parameters” on page 64.
For information on operators in general, see Operators in the Language Quick
Reference.
Piecewise-linear functions
Describes the use of piecewise-linear functions in OPL.
Piecewise-linear functions are important in many applications. They are often
specified by giving a set of slopes, a set of breakpoints at which the slopes change,
and the value of the functions at a given point. Consider, for instance, a
Chapter 2. OPL, the modeling language 43
transportation problem in which the transportation cost between two locations o
and ddepends on the size of the shipment ship[o][d]. The piecewise-linear
expression
piecewise{10 -> 100;20 -> 200;40}(0,0) ship[o][d];
describes the piecewise-linear function of ship[o,d] depicted in Figure 1 The
function has slopes 10, 20, and 40, breakpoints 100 and 200, and evaluates to 0at
point 0.
In other words, the piecewise-linear expression is equivalent to the expression:
10 * ship[o][d]
when
ship[o,d] <= 100
equivalent to
(10 * 100) + 20 * (ship[o][d] - 100)
when
100 <= ship[o][d] <= 200
and equivalent to
(10 * 100) + (20 * 200) + 40 * (ship[o][d] - 200)
otherwise.
By default, OPL assumes that a piecewise-linear function evaluates to zero at the
origin, so that the above piecewise-linear function could actually be written as
piecewise{10 -> 100;20 -> 200;40} ship[o][d];
The above piecewise-linear function has a fixed number of pieces, but OPL also
allows generic pieces. The number of pieces may then depend on the input data,
as in
Figure 1. A Piecewise-linear function.
44 OPL Language Reference Manual
piecewise(i in 1..n) {
slope[i] -> breakpoint[i];
slope[n+1];
} ship[o][d];
If breakpoint[1] ≥ 0 this piecewise-linear function is equivalent to
slope[1] * ship[o][d]
when
ship[o][d] <= breakpoint[1]
It is equivalent to
when
breakpoint[k-1] < ship[o][d] <= breakpoint [k] (1<k<=n)
and equivalent to
otherwise.
Note that there may be several generic pieces in piecewise-linear functions. It is
important to stress that breakpoints and slopes in piecewise-linear functions must
always be grounded by a point on the piecewise-linear function. Such a point
(called an anchor point) uniquely defines the function. Also, the breakpoints must
be strictly increasing.
To sort your model data for this purpose, use sorted sets, as explained in “Sorted
and ordered sets” on page 13.
Section Piecewise linear programming in the Language User’s Manual discusses
piecewise-linear functions applied to an inventory problem.
Discontinuous piecewise-linear functions
Describes the use of discontinuous piecewise-linear functions in OPL.
OPL also allows you to write discontinuous piecewise-linear functions. This is the
case when, in the syntax of a piecewise-linear function with slopes and
breakpoints, two successive breakpoints are identical and the value associated with
the second one is considered to be a “step” instead of a “slope”. The CPLEX and
the CP Optimizer engines behave differently with respect to what limit they
consider as the discontinuity value. Because CPLEX allows either of these limits,
note that the anchor point used to ground the breakpoints and slopes must not
reside at the discontinuity. Otherwise, the piecewise-linear function would not be
uniquely defined.
Chapter 2. OPL, the modeling language 45
Behavior with the CPLEX engine
Example 1: the sign function
The following piecewise function:
piecewise{0-&gt;0; 2-&gt;0; 0}(1,1) x;
has a slope of 0 up to breakpoint 0, then a step of 2 at this breakpoint, then a slope
of 0. It takes the value 1 at point 1. This piecewise represents the function sign()
which returns the sign (1 or -1) of its argument, as represented in Figure 1 below.
Then this model
dvar float x;
dvar float signx;
dvar float y;
dvar float signy;
maximize x;
subject to {
x==2;
signx == piecewise{0->0; 2->0; 0}(1,1) x;
y==-2;
signy == piecewise{0->0; 2->0; 0}(1,1) y;
}
gives the following output:
Figure 2. The discontinuous piecewise-linear function sign()
46 OPL Language Reference Manual
Final solution with objective 2.0000:
x = 2.0000;
signx = 1.0000;
y = -2.0000;
signy = -1.0000;
Figure 2 on page 46 shows that the value of the sign function at the breakpoint is
either -1 (on the left-hand slope) or 1(on the right-hand slope).
For example, this model takes this into account and sets the constraint x==y; on
both values.
dvar float x;
dvar float signx;
dvar float y;
dvar float signy;
maximize signx-signy;
subject to {
x==y;
signx == piecewise{0->0; 2->0; 0}(1,1) x;
signy == piecewise{0->0; 2->0; 0}(1,1) y;
}
This model solves with the following output:
Final solution with objective 2:
signx = 1;
signy = -1;
x=0;
y=0;
Example 2: discontinuous cost
The following piecewise function
piecewise{0-&gt;0; 10-&gt;0; 0-&gt;10; 5-&gt;10; 0-&gt;20; 5-&gt;20; 0} (5,10) unit;
represents a discontinuous cost.
This function is illustrated in Figure 2, Discontinuous costs, for the values
summarized in Table 1.
Table 5. A discontinuous cost function
Values of Unit Cost
&lt0 0
0to10 10
10 to 20 15
&gt;20 20
Chapter 2. OPL, the modeling language 47
Different behavior with the CP Optimizer engine
Consider the following model:
//using CP;
dvar int x in -10..10;
dvar int signx;
dvar int y in -10..10;
dvar int signy;
maximize signx-signy;
subject to {
x==y;
signx == piecewise{0-&gt;0; 2-&gt;0; 0}(1,1) x;
signy == piecewise{0-&gt;0; 2->0; 0}(1,1) y;
}
execute
{
writeln(signx-signy);
}
Depending on which solving engine you write for, you get a different result
because CPLEX and CP Optimizer do not handle limit values in the same way.
vIf you comment out the using CP; statement, the model is solved by the CPLEX
engine and the result is 2because CPLEX handles symmetry in such a way that
it interprets either limit as the discontinuity value.
vHowever, if you uncomment the using CP; statement, the model is solved by
the CP Optimizer engine and the result is 0because CP Optimizer considers the
left limit as the discontinuity value.
Set expressions
Describes the use of set expressions in OPL.
Figure 3. Discontinuous costs
48 OPL Language Reference Manual
Set data can be initialized by set expressions, as mentioned in “Data types” on
page 4. This section describes how these expressions are constructed and what
functions are defined over sets.
v“Construction of set expressions”
v“Functions for sets”
Construction of set expressions
Set expressions are constructed from previously defined sets and the set operations
union,inter,diff, and symdiff. For instance:
{int} s1 = {1,2,3};
{int} s2 = {1,4,5};
{int}i=s1inter s2;
{int}u=s1union s2;
{int}d=s1diff s2;
{int} sd = s1 symdiff s2;
initializes ito {1},uto {1,2,3,4,5},dto {2,3}, and sd to {2,3,4,5}. In addition,
set expressions can be constructed from ranges. For instance, the excerpt
{int} s = asSet(1..10);
initializes sto the finite set {1,2,...,10}
Important:
The range 1..10 takes constant space, while the set stakes space proportional to
the number of elements in the range.
Functions for sets
Table 6 shows the functions available over sets. These methods apply to all sets,
including sets of tuples and other complex sets. In this section, we assume that:
vSis the set 3, 6, 7, 9
vitem is an element of S
vnis an integer number
Table 6. Functions over sets
Function Description
card card(S) returns the size of S, that is, the
number of items.
ord ord(S,item) returns the position of item in
S. Positions start at 0and ord(S,item)
produces an execution error if item is not in
S.
Example: ord(S,6) evaluates to 1and
ord(S,9) to 3.
The order of items in an explicit set is by
order of appearance in the initialization and
is implementation-dependent when the sets
are the results of a set operation.
first first(S) returns the first item in S,3in this
example.
Chapter 2. OPL, the modeling language 49
Table 6. Functions over sets (continued)
Function Description
item item(S,n) returns the n-th item in set S.
Counting starts from 0. This is equivalent to
next(first(S),n)
Example: item(S,1) = 6
last last(S) returns the last item in S,9in this
example.
next next(S,item) returns the item in S that
comes after item and produces an execution
error if item is the last item.
Example: next(S,3) = 6
next(S,item,n) returns the n-th next item.
next(S,item) is equivalent to
next(S,item,1).
nextc A circular version of next.nextC(S,item)
returns the first item in Sif item is the last
item.
Example: nextc(S,9) = 3
nextc(S,item,n) returns the n-th circular
next item. nextc(S,item) is equivalent to
nextc(S,item,1).
prev prev(S,item) returns the item in Sthat
comes before item and produces an
execution error if item is the first item.
Example: prev(S,6) = 3
prev(S,item,n) returns the n-th previous
item. prev(S,item) is equivalent to
prev(S,item,1).
prevc A circular version of prev.prevc(S,item)
returns the last item in Sif item is the first
item.
Example: prev(S,3) = 9
prevc(S,item,n) returns the n-th circular
previous item. prevc(S,item) is equivalent
to prevc(S,item,1).
Boolean expressions
Describes the use of Boolean expressions in OPL.
Boolean expressions can have various operand types in OPL. They are constructed
in different ways:
vfrom integer expressions using the traditional relational operators ==,!= (not
equal), >=,>,<, and <=.
vfrom float expressions using the same relational operators.
vfrom string expressions and support the same operators as well.
50 OPL Language Reference Manual
For convenience, OPL offers a range expression to express special combinations for
constraints.
They are of the form
a op1 x op2 b
where
vop1 and op2 are either of the relational operators <= or <
vaand bare boundary expressions which need to be ground
vxis an expression
Those expressions are equivalent to
a op1 x; x op2 b
which is itself equivalent to
aop1x&&xop2b
Constraints
Specifies the constraints supported by OPL and discusses various subclasses of
constraints to illustrate the support available for modeling combinatorial
optimization applications.
Introduction
Constraints are a subset of Boolean expressions.
The availability of certain constraints depends on their context. The contexts can
be:
vData initialization when declared data is assigned
vOptimization model
–aconstraints block
–asubject to block
vAn expression that filters an iteration for aggregation or generation.
Using constraints
Explains how to apply a constraint to a decision variable and, possibly,
conditionalize it. Also explains why constraints are identified by a label, and why
constraints are used for filtering purposes.
Modeling constraints
Shows how to define modeling constraints in OPL.
Constraints passed to the algorithm, and which as such define the optimization
problem, usually apply to decision variables; that is, they are Boolean expressions
of some decision variables. To be taken into account by the solving algorithm,
constraints must be stated using the optimization instruction:
constraints
or
subject to
Chapter 2. OPL, the modeling language 51
as shown in the example below.
Stating constraints by means of an optimization instruction
minimize
sum(p in Products) (insideCost[p]*inside[p] + outsideCost[p]*outside[p]);
subject to {
forall(r in Resources)
sum(p in Products) consumption[p,r] * inside[p] <= capacity[r];
forall(p in Products)
inside[p] + outside[p] >= demand[p];
}
Note:
1. Optimization instructions require an objective function of type integer or float.
2. That objective function must be defined before the constraints. Otherwise, a
warning message is displayed.
Conditional constraints
Shows how to define conditional constraints in OPL.
If-then-else statements make it possible to state constraints conditionally, as in:
if(d>1){
abs(freq[f] - freq[g]) >= d;
} else {
freq[f] == freq[g];
}
Conditions in if-else statements must be ground; that is, they must not contain
decision variables. Also, the if block should not contain forall statements such as:
if (..) {
forall(...)
...
}
Implications of constraints can be used instead when conditions contain decision
variables. Conditionals can also be used in OPL to make different choices
according to the truth value of a condition.
Filtering with constraints
Describes the process of using constraints to filter decision variables or aggregates.
In addition to applying constraints to decision variables, you can also create
constraints on formal parameters to filter aggregates, like this:
Filtering with constraints
// The cities where we are doing business
{string} cities={"Paris","Berlin","Washington","Rio"};
{string} EuropeanMainCapitals = {"London","Paris","Berlin","Madrid","Roma"};
// Should we expand business in this city ?
dvar boolean x[cities];
// We want to expand business in Europe
maximize sum(c in cities: c in EuropeanMainCapitals) x[c];
subject to
{
// We can expand business in 2 cities
sum(c in cities) x[c]<=2;
52 OPL Language Reference Manual
}
{string} expanded_cities = {c|cincities : x[c]==1};
execute
{
for(c in expanded_cities) writeln("We should expand business in ",c);
}
The result is:
We should expand business in Paris
We should expand business in Berlin
In this context, filtering means placing a condition on an iteration to limit the
iteration loop. See “Formal parameters” on page 64 for more examples.
Constraint labels
Explains why label constraints, the benefits, costs, and limitations, how to label
constraints, how to use indexed labels, and how to deal with compatibility
between constraint names and labels.
Why label constraints?
Explains why attaching labels to your constraints is the recommended practice.
You can identify constraints by attaching labels to them. It is the recommended
practice but it has a performance cost.
Benefits
vConstraint labels enable you to benefit from the expand feature in the IDE
Problem Browser to find which constraints are tight in a given application or to
find dual variable values in linear programs. See Understanding the Problem
Browser in Getting Started with the IDE.
vYou can access the slack and dual values for labeled constraints when a solution
is available. See the class IloConstraint in the Reference Manual of IBM ILOG
Script Extensions for OPL.
vOnly labeled constraints are considered by the relaxation and conflict search
process in infeasible models (see Relaxing infeasible models in IDE Tutorials).
Cost
However, labeling constraints has a performance and memory cost which can be
significant, especially when a tuple pattern is used as the index. Therefore, you are
encouraged to not use labels for large models or, if you do, at least use tuple
indices instead of tuple patterns.
More precisely, constraint labels are used in three cases: IDE expand actions, in
slack and dual values with solutions, and with relaxation and conflict detection. If
you do not need these three use cases, you should get rid of the label to speed up
the execution and lower memory consumption.
Labeling constraints
Shows how to label constraints in OPL.
Chapter 2. OPL, the modeling language 53
Procedure
Type the character string you want, followed by the colon (:) sign, before the
constraint you want to label, as shown in Labelling constraints (production.mod).
If you used to declare constraint names in your existing OPL models, see
“Compatibility between constraint names and labels” on page 57.
Note:
A constraint label or name cannot start with a number.
Labeling constraints (production.mod)
minimize
sum( p in Products )
( InsideCost[p] * Inside[p] + OutsideCost[p] * Outside[p] );
subject to {
forall( r in Resources )
ctCapacity:
sum( p in Products )
Consumption[p][r] * Inside[p] <= Capacity[r];
forall(p in Products)
ctDemand:
Inside[p] + Outside[p] >= Demand[p];
}
This is equivalent to “Stating constraints by means of an optimization instruction”
on page 52. The only difference is that the constraint on the production capacity
has been labeled
ctCapacity:
and the constraint on the demand of products has been labeled
ctDemand:
These labels can be used to display the data.
Using indexed labels
Shows how to use indexed labels for your OPL constraints.
About this task
In some cases, it is more convenient to use indexed labels. Indexed labels enable
you to control how a constraint is assigned to an array item.
The example below, Labeling constraints with indexed labels shows that the
transp2.mod model identifies constraints using indexed labels following this
syntax:
constraint ctDemand[Products];
...
ctDemand[p]:...
Indexed labels on constraints (transp2.mod)
forall( p in Products,dinDest[p] )
ctDemand[p][d]:
sum( o in Orig[p] )
Trans[< p,o,d >] == Demand[<p,d>];
ctCapacity: forall(o,dinCities )
sum( <p,o,d> in Routes )
Trans[<p,o,d>] <= Capacity;
54 OPL Language Reference Manual
A case where you need indexed labels to reduce memory overhead is when you
use forall iterations with variable sizes, as shown in the following code.
forall iterations with variable sizes
forall( p in Products,oinOrig[p] )
In this forall example, the second formal parameter oiterates on sets of
potentially different sizes, depending on the value of the formal parameter p.
To use indexed labels:
Procedure
1. Declare the constraint array that will receive the labeled constraints.
constraint ctSupply[Products][Cities];
2. Add the indexing expressions to the label.
ctSupply[p][o]:
Example
The full code extract is shown here.
Labeling constraints with indexed labels
constraint ctSupply[Products][Cities];
constraint ctDemand[Products][Cities];
minimize
sum(l in Routes) Cost[l] * Trans[l];
subject to {
forall( p in Products,oinOrig[p] )
ctSupply[p][o]:
sum( d in Dest[p] )
Trans[< p,o,d >] == Supply[<p,o>];
forall( p in Products,dinDest[p] )
ctDemand[p][d]:
sum( o in Orig[p] )
Trans[< p,o,d >] == Demand[<p,d>];
Labeled assert statements
Shows how to label assertions in OPL.
Assertions can be labeled. When you label a constraint that is part of an assert
statement, and if the assertion fails, the context of the failing assertions appears in
the Issues output window. For example:
{int} vals = {1, 2, 3};
assert forall(i in vals) ct:i<2;
Chapter 2. OPL, the modeling language 55
Limitations on constraint labeling
Explains what constraints cannot be labeled or at what cost they can be.
Not all constraints can be labeled. Limitations exist with respect to forall
statements and to a variable size indexer.
Labels and forall statements
You can label only constraints that are not nested within a forall statement (leaf
constraints). However, you can label a forall constraint, if it is at the root level of
constraints. For example, in the code sample Constraint label within forall
statement, the constraint
ct1: forall(i in r1) forall(j in r2) X[i][j] <= i+j;
can also be written
forall(i in r1) forall(j in r2) ct1: X[i][j] <= i+j;
In both cases, the model executes correctly. However, if you execute:
forall(i in r1) ct1: forall(j in r2) X[i][j] <= i+j;
the IDE reports “Element "ct1" has never been used” and the constraint does not
appear in the Problem Browser.
Constraint label within forall statement
range r1 = 1..2;
range r2 = 1..3;
dvar int X[r1][r2] in 0..5;
constraints {
ct1: forall(i in r1) forall(j in r2) X[i][j] <= i+j;
}
Labels and a variable size indexer
Constraint labels with a variable size indexer are not allowed. For example, this
model generates an error message.
tuple RangeTuple
{
int i;
int j;
string k;
};
{RangeTuple} RT = {<1, 2, "bla">};
Figure 4. Labeled assert
56 OPL Language Reference Manual
minimize 1;
subject to
{
forall(<p1, p2, p3> in RT)
forall(i in p1..p2)
rangeLabel:
1==1;
}
Write the following code instead.
tuple RangeTuple
{
int i;
int j;
string k;
};
{RangeTuple} RT = {<1, 2, "bla">};
{int} s={1,2};
constraint rangeLabel[RT][s];
minimize 1;
subject to
{
forall(<p1, p2, p3> in RT)
forall(i in p1..p2)
rangeLabel[<p1, p2, p3>,i]:
1==1;
}
Note the difference in rangeLabel.
Compatibility between constraint names and labels
Explains compatibility issues for constraint names between OPL 4.x and
subsequent versions.
OPL 4.x constraint names are deprecated in OPL 5.0 and later. They are still
supported to maintain the compatibility of your OPL 4.x models and automatically
implemented as labels internally, but they will be removed in future versions of
IBM ILOG OPL. It is therefore strongly recommended that:
vyou label constraints that are currently neither named nor labeled, as shown in
“Labeling constraints” on page 53,
vyou change possible existing constraint names in your models to labels.
Compare Deprecated constraint names with Labeling constraints
(production.mod). You do not need to previously declare the label as was the
case with constraint names and you use the colon (:) sign instead of the equal
(=) sign.
Deprecated constraint names
constraint capacityCons[r];
constraint demandCons[p];
minimize
sum(p in Products) (insideCost[p]*inside[p] + outsideCost[p]*outside[p]);
subject to {
forall(r in Resources)
Chapter 2. OPL, the modeling language 57
capacityCons[r]= sum(p in Products) consumption[p,r] * inside[p] <= capacity[r];
forall(p in Products)
demandCons[p]= inside[p] + outside[p] >= demand[p];
}
Types of constraints
Classifies constraints according to their operand type.
Float constraints
Float constraints are constraints involving float decision variables.
When modeled in OPL, they are restricted to be linear or piecewise linear. OPL has
efficient algorithms for solving linear, piecewise linear, quadratic, or logical
constraints, but in general these algorithms do not apply to nonlinear problems.
Note that the linearity requirement precludes the use of relations with variables in
constraints and the use of non-ground expressions as indices of float arrays. In
addition, operators !=,<, and >are not allowed for float constraints. However,
integers and integer variables may occur in a float constraint, provided that the
constraint remains linear or quadratic. OPL supports all the expressions supported
by CPLEX, provided that the constraint is of one of the types described at the
beginning of this paragraph.
Discrete constraints
Discrete constraints are arbitrary Boolean expressions with integer operands,
possibly containing variables.
In OPL, these constraints must be well-typed, but no restrictions are imposed on
them. It is, however, useful to review subclasses of these constraints to illustrate
the functionalities of OPL. MP models, solved by the CPLEX engine, can contain
only basic constraints (see the section Basic constraints below. Logical constraints
can be set by filtering of forall and sum constraints.
Basic constraints
Basic discrete constraints are constructed from discrete data, discrete variables, and
the arithmetic operators and functions defined in “Expressions” on page 41. For
instance, the excerpt
range r=1..5;
dvar int x[1..5] in 0..10;
dvar int obj;
maximize obj;
subject to
{
obj==sum(ordered i,j in r) abs(x[i]-x[j]);
forall(ordered i,j in r) abs(x[i]-x[j])>=1;
}
generates distance constraints between integer variables.
Note that the following code creates an error because IBM ILOG CPLEX does not
accept non linear constraints.
dvar int+ X in 0..1000;
minimize X;
subject to {
Xmod7==0;
};
58 OPL Language Reference Manual
String constraints
Describes the use of string constraints in OPL.
String constraints cannot be used on decision variables or added to the model to
be solved. They can only be used on indexers to filter aggregates (see “Filtering
with constraints” on page 52). For example:
{string} s = {"a", "b"};
dvar int x[s] in 0..10;
minimize sum(i in s) x[i];
subject to {
forall(i ins:i!="a")
x[i] >= 5;
}
Implicit constraints
Explains that implicit constraints may imply infeasibility.
Implicit constraints are implied by operators. For example:
using CP;
int a[1..4];
dvar int x in 1..5;
maximize a[x];
subject to
{
x==5;
}
makes the model infeasible even though
using CP;
int a[1..4];
dvar int x in 1..5;
subject to
{
x==5;
}
is feasible.
Logical constraints for CPLEX
Describes the use of logical constraints in OPL.
Logical constraints are one particular kind of discrete or numerical constraints.
OPL and CPLEX can translate logical constraints automatically into their
transformed equivalent that the discrete (MIP) or continuous (LP) optimizers of
IBM ILOG CPLEX can process efficiently. This section describes all the available
logical constraints, as well as the logical expressions that can be used in logical
constraints. Logical constraints are available in constraint programming models
without linearization.
For an example of how OPL uses logical constraints, see Tutorial: Using CPLEX
logical constraints in the Language User’s Manual.
In this section, you will learn:
v“What are logical constraints?” on page 60
Chapter 2. OPL, the modeling language 59
v“What can be extracted from a model with logical constraints?”
v“Which nonlinear expressions can be extracted?”
v“Logical constraints for counting” on page 61
v“How are logical constraints extracted?” on page 61
What are logical constraints?
For IBM ILOG CPLEX, a logical constraint combines linear constraints by means of
logical operators, such as logical-and, logical-or, implication, negation (not),
conditional statements (if ... then ...) to express complex relations between
linear constraints. IBM ILOG CPLEX can also handle certain logical expressions
appearing within a linear constraint. One such logical expression is the minimum
of a set of variables. Another such logical expression is the absolute value of a
variable. There’s more about logical expressions in “Which nonlinear expressions
can be extracted?.”
What can be extracted from a model with logical constraints?
The table below lists the logical constraints that CPLEX can extract.
Symbol Meaning
&& Logical AND
| | Logical OR
! Logical NOT
=> Imply
!= Different from
== Equivalence
All those constructs accept as their arguments other linear constraints or logical
constraints, so you can combine linear constraints with logical constraints in
complicated expressions in your application.
Note:
With the C++ API you should use the symbol <= to express implication.
Which nonlinear expressions can be extracted?
Some expressions are easily recognized as nonlinear, for example, a function such
as
x^2+y^2=1
However, other nonlinearities are less obvious, such as absolute value as a
function. In a very real sense, MIP is a class of nonlinearly constrained problems
because the integrality restriction destroys the property of convexity which any
linear constraints otherwise might possess. Because of that characteristic, certain
(although not all) nonlinearities are capable of being converted to a MIP
formulation, and thus can be solved by IBM ILOG CPLEX. The following nonlinear
expressions are accepted in an OPL model:
vmin and minl : the minimum of several numeric expressions
vmax and maxl : the maximum of several numeric expressions
vabs : the absolute value of a numeric expression
60 OPL Language Reference Manual
vpiecewise : the piecewise linear combination of a numeric expression
vA linear constraint can appear as a term in a logical constraint.
In fact, ranges containing logical expressions can, in turn, appear in logical
constraints. It is important to note here that only linear constraints can appear as
arguments of logical constraints extracted by IBM ILOG CPLEX. That is, quadratic
constraints are not handled in logical constraints. Similarly, quadratic terms can not
appear as arguments of logical expressions such as min,max,abs, and piecewise.
Logical constraints for counting
In many cases it is even unnecessary to allocate binary variables explicitly in order
to gain the benefit of linear constraints within logical expressions. For example,
optimizing how many items appear in a solution is often an issue in practical
problems. Questions of counting (how many?) can be represented formally as
cardinality constraints. Suppose that your application includes three variables, each
representing a quantity of one of three products, and assume further that a good
solution to the problem means that the quantity of at least two of the three
products must be greater than 20. Then you can represent that idea in your
application, like this:
(x[0] >= 20) + (x[1] >= 20) + (x[2] >= 20) >= 2;
How are logical constraints extracted?
Logical constraints are transformed automatically into equivalent linear
formulations when they are extracted by an IBM ILOG CPLEX algorithm. This
transformation involves automatic creation by IBM ILOG CPLEX of new variables
and constraints. For more details on this transformation, refer to the IBM ILOG
CPLEX documentation.
Constraints available in constraint programming
Lists the types of constraints and expressions available in OPL for constraint
programming models. See the OPL Language Quick Reference and the CP Optimizer
User’s Manual for further details.
v“Arithmetic constraints and expressions”
v“Logical constraints for CP” on page 62
v“Element constraints” on page 62
v“Compatibility constraints” on page 63
v“Specialized constraints” on page 63
Arithmetic constraints and expressions
The following arithmetic constraints and expressions are available for OPL CP
models. The references point to the OPL Language Quick Reference.
vArithmetic operations
– addition
– subtraction
– multiplication
– scalar products
– integer division
– floating-point division
– modular arithmetic
Chapter 2. OPL, the modeling language 61
vArithmetic expressions for use in constraints.
– standard deviation: see standardDeviation
– minimum: see min
– maximum: see max
– counting: see count
– absolute value: see abs
– element or index: see element
vArithmetic constraints
– equal to
– not equal to
– strictly less than
– strictly greater than
– less than or equal to
– greater than or equal to
Logical constraints for CP
Logical operators, such as logical-and, logical-or, implication, negation (not),
difference, and equivalence are available in constraint programming.
The following logical constraints are available for OPL CP models.
Symbol Meaning
&& Logical AND
| | Logical OR
! Logical NOT
=> Imply
!= Different from
== Equivalence
Note:
With the C++ API you should use the symbol <= to express implication.
Element constraints
The following subscripting operators return constrained integer or floating-point
expressions, also known as element expressions.
vA[X], where A is an array of integer values (with one or several dimensions) and
X is a dexpr int, returns a dexpr int such that: when X is fixed to the value i, the
value of the expression is A[i].
vA[X], where A is an array of floating-point values (with one or several
dimensions) and X is a dexpr int, returns a dexpr float such that: when X is
fixed to the value i, the value of the expression is A[i].
More generally, the domain of the expression is the set of values A[i], where i is in
the domain of X.
In both of the above cases, the expression can include several indices (A[X][k],
A[X][Y], A[X][k]{Y], ...).
62 OPL Language Reference Manual
vFor constant indices, all the usual index types are available (int, float, string,
tuple).
vFor a decision expression index, only the integer type is available (dvar int or
dexpr int).
Compatibility constraints
The CP Optimizer engine supports allowed and forbidden assignments for OPL CP
models. See allowedAssignments and forbiddenAssignments in the Language Quick
Reference.
Scheduling constraints
This section lists the scheduling constraints available for OPL CP scheduling
models and provides links to the reference documentation for these constraints. For
a more detailed description of these constraints, please refer to the “Scheduling” on
page 68 section of this manual.
The following constraints are available for OPL CP scheduling models:
vPrecedence constraint: endAtEnd
vPrecedence constraint: endAtStart
vPrecedence constraint: endBeforeEnd
vPrecedence constraint: endBeforeStart
vPrecedence constraint: startAtEnd
vPrecedence constraint: startAtStart
vPrecedence constraint: startBeforeEnd
vPrecedence constraint: startBeforeStart
vInterval grouping constraint: alternative
vInterval grouping constraint: span
vInterval grouping constraint: synchronize
vInterval presence constraint: presenceOf
vSequence constraint: first (scheduling)
vSequence constraint: last (scheduling)
vSequence constraint: before
vSequence constraint: prev
vSequence constraint: noOverlap
vCumulative or state function constraint: <= (scheduling), alwaysIn
vState function constraint: alwaysConstant
vState function constraint: alwaysEqual
vState function constraint: alwaysNoState
Specialized constraints
The CP Optimizer engine also accepts some powerful combinatorial constraints
known as specialized constraints. For these constraints, some powerful propagation
algorithms are used to reduced the decision variable domains.
vallDifferent : constrains variables within a dvar array to all take different
values
vallMinDistance : constrains variables within a dvar array to all take values that
are one-to-one different by at least a given gap
Chapter 2. OPL, the modeling language 63
vinverse : takes two arrays of integer variables that must be indexed by an
integer and be one-dimensional
vlex : states that the first array of variables is less than or equal to the second
array of variables in the alphabetical order
vpack : represents some simple but powerful one-dimensional packing constraint
See the individual entries under OPL functions in Language Quick Reference for a
complete description of each constraint.
Limitations on constraints
There is a limitation on certain types of constraints.
You cannot change the bound of a top level inequality constraint unless that bound
is a literal expression, such as x≥13or x≤13.
As a workaround, you can declare your constraint either by using a range syntax,
or by putting the inequality inside a forall statement.
Instead of
ct0: x ≤ a[0];
you can write
ct1: -infinity≤x≤a[0];
or
forall(i in 0..0)
ct2: x ≤ a[0];
and change the upper (or lower) bound of the constraint with a script instruction
such as:
ct1.UB = 1313;
or
ct2[0].UB = 1313;
Formal parameters
Describes basic formal parameters, tuples of parameters, and filtering in tuples of
parameters.
Basic formal parameters
Formal parameters play a fundamental role in OPL; they are used in aggregate
operators, generic sets, and forall statements.
The simplest formal parameter has the form
pinS
where pis the formal parameter and Sis the set from which ptakes its values.
The set Scan be:
van integer range, as in
64 OPL Language Reference Manual
int n=6;
int s = sum(i in 1..n) i*i;
va string set, as in
{string} Products ={"car","truck"};
float cost[Products] =[12000,10000];
float maxCost = max(p in Products) cost[p];
vor a tuple set, as in
{string} Cities = { "Paris", "London", "Berlin" };
tuple Connection
{
string orig;
string dest;
}
{Connection} connections = { <"Paris","Berlin">,<"Paris","London">};
float cost[connections] = [ 1000, 2000 ];
float maxCost= max(r in connections) cost[r];
If you need to filter the range of the formal parameters using conditions, the
formal parameter then takes the form
p in S : filtering condition
and assigns to pall elements of Saccording to the filter applied.
For instance, in the excerpt
int n=8;
dvar int a[1..n][1..n];
subject to
{
forall(i in 1..8)
forall(j in 1..8:i<j)
a[i][j] >= 0;
}
the constraint a[i][j] >= 0 is modeled for all iand jsuch that 1≤i<j≤8.
Note:
OPL does not support aggregates in filter expressions. For example:
vFor an expression such as:
{int} notFirst = {i|iin1..10 : card({j | <i,j> in pairs}) == 0};
the type check displays the error: "Aggregate set is currently not supported for
filter expressions."
vFor an expression such as:
{int} notFirst = {i|iin1..10 : sum(<i,j>in pairs) i == 0};
the type check displays the error: "Aggregate sum is currently not supported for
filter expressions."
Several parameters can often be combined together to produce more compact
statements. For instance, the declaration
int s = sum(i,j in 1..n:i<j)i*j;
is equivalent to
int s = sum(i in 1..n) sum(j in 1..n:i<j)i*j;
which is less readable.
Chapter 2. OPL, the modeling language 65
The declaration
int s = sum(i in 1..n, j in 1..m) i*j;
is equivalent to
int s = sum(i in 1..n) sum(j in 1..m) i*j;
These parameters can, of course, be subject to filtering conditions. The excerpt
forall(i,j in 1..n:i<j)
a[i][j] >= 0;
is equivalent to
forall(i in 1..n, j in 1..n : i<j)
a[i][j] >= 0;
Here is an even more compact form:
forall(ordered i,j in 1..n)
a[i][j] >= 0;
Indeed, in many applications one is interested, given a set S, in stating filters or
conditions over all pairs (i, j) of elements of Ssatisfying i<jin the ordering
associated with S. In this excerpt
{T} S = ...;
forall(ordered s, t in S)...;
forall(s in S, t in S: ord(S,s) < ord(S,t)) ...
the first forall line is equivalent to the second one and illustrates the functionality
ordered, often useful in practical applications. T can be one of the types int,float,
string, or a tuple type.
Important:
This ordering does not refer to the ordering associated with type T but to the order
of the items within the set.
Tuples of parameters
OPL allows tuples of formal parameters to be created in aggregate operators,
forall statements, and generic sets.
The code sample below states the precedence constraints between tasks. The
constraint declaration requires explicit accesses to the fields of the tuple to state the
constraints. In addition, the field before is accessed twice. An alternative way to
state the same constraint is to use a tuple of formal parameters, as shown in the
last line of the code sample, precluding the need to access the tuple fields
explicitly. The tuple <p in Prec> in the forall quantifier contains two components
that are successively given the values of the fields of each tuple in Prec.
Tuple of formal parameters
int minTime=7*60;
int maxTime=9*60;
{string} Tasks = { "Make dinner","Have dinner","Clean post dinner" };
tuple Precedence {
string pre;
string post;
}
{Precedence} Prec = {
66 OPL Language Reference Manual
<"Make dinner","Have dinner">,
<"Have dinner","Clean post dinner">
};
int Duration[Tasks]= [20,60,10];
dvar int Start[Tasks] in minTime..maxTime;
subject to {
forall(p in Prec) Start[p.post] >= Start[p.pre] + Duration[p.pre];
}
More generally, an expression
pinS
where Sis a set of tuples containing nfields, can be replaced by a formal
parameter expression
<p1,...,pn> in S
that contains nformal parameters. Each time a tuple ris selected from S, its fields
are assigned to the corresponding formal parameters. This functionality is often
useful in producing more readable models.
Filtering in tuples of parameters
OPL enables simple equality constraints to be factorized inside tuples, which is
important in obtaining more readable and efficient models. In this context, slicing
refers to nested iterations with filtering conditions.
Consider, for instance, a transportation problem where products must be shipped
from one set of cities to another set of cities. The model may include a constraint
specifying that the total shipments for all products transported along a connection
may not exceed a specified limit. This can be expressed by a constraint
Explicit slicing
forall(c in connections)
sum(<p,co> in routes: c == co) trans[<p,c>] <= limit;
This constraint states that the total products shipped along each connection cis not
greater than limit. OPL must scan the entire set routes to select the tuples
involving each connection. In this example, the expression c==co is used to make
slicing explicit.
The constraint would be stated equivalently as follows:
Implicit slicing
forall(c in connections)
sum(<p,c> in routes) trans[<p,c>] <= limit;
In this constraint, the tuple <p,c> contains one new parameter pand uses the
previously defined parameter c. Since the value of cis known, OPL uses it to
index the set routes, avoiding a complete scan of the set routes. In this example,
slicing is said to be implicit because the formal parameter cis used to declare
iteration in both the forall and sum loops. You can also use a constant as a tuple
item, for example <p,2>, for implicit slicing.
Chapter 2. OPL, the modeling language 67
In OPL 4.0 and later versions, models tend to be more readable when explicit
slicing is used. Besides, there is no performance advantage in using implicit slicing
over explicit slicing.
More about implicit slicing
You should be aware that the following statement:
int array[i in set1] = ((sum(i in set2) 1 >= 1) ? 1:0);
is exactly equivalent to
int array[i in set1] = ((sum(j in set2) 1 >= 1) ? 1:0);
that is, the two “i” on either side of the “equal” sign =, in the first statement, are
not linked. This is called scope hiding because the second “i” hides the first one in a
nested scope.
In contrast, this statement
int array[<i,j> in set1] = ((sum(<i,j> in set2) 1 >= 1) ? 1:0);
codes implicit slicing, which is equivalent to:
int array[i in set1] = ((sum(j in set2 : j==i) 1 >= 1) ? 1:0);
In other words, there is no implicit slicing outside tuple patterns.
See also Modeling tips in the Language User’s Manual.
Scheduling
Describes how to model scheduling problems in OPL.
Introduction
OPL and CP Optimizer introduce a set of modeling features for applications that
deal with scheduling over time.
In OPL and CP Optimizer, time points are represented as integers, but the possible
very wide range of time points means that time is effectively continuous. A
consequence of scheduling over effectively continuous time is that the evolution of
some known quantities over time (for instance the instantaneous efficiency/speed
of a resource or the earliness/tardiness cost for finishing an activity at a given date
t) needs to be compactly represented in the model. To that end, CP Optimizer
provides the notion of piecewise linear and stepwise functions.
Most scheduling applications consist in scheduling in time some activities, tasks or
operations that have a start and an end time. In CP Optimizer, this type of decision
variable is captured by the notion of the interval variable. Several types of
constraints are expressed on and between interval variables:
vto limit the possible positions of an interval variable (forbidden start/end or
“extent” values)
vto specify precedence relations between two interval variables
vto relate the position of an interval variable with one of a set of interval
variables (spanning, synchronization, alternative).
An important characteristic of scheduling problems is that time intervals may be
optional, and whether to execute a time-interval or not is a possible decision
68 OPL Language Reference Manual
variable. In CP Optimizer, this is captured by the notion of a boolean presence
status associated with each interval variable. Logical relations can be expressed
between the presence of interval variables, for example to state that whenever
interval ais present then interval bmust also be present.
Another aspect of scheduling is the allocation of scarce resources to time intervals.
The evolution of a resource over time can be modelled by two types of variables:
vThe evolution of a disjunctive resource over time can be described by the
sequence of intervals that represent the activities executing on the resource. For
that, CP Optimizer introduces the notion of an interval sequence variable.
Constraints and expressions are available to control the sequencing of a set of
interval variables.
vThe evolution of a cumulative resource often needs a description of how the
accumulated usage of the resource evolves over time. For that purpose, CP
Optimizer provides the concept of the cumulative function expression that can be
used to constrain the evolution of the resource usage over time.
vThe evolution of a resource of infinite capacity, the state of which can vary over
time, is captured in CP Optimizer by the notion of the state function. The
dynamic evolution of a state function can be controlled with the notion of
transition distance, and constraints are available for specifying conditions on the
state function that must be satisfied during fixed or variable intervals.
Some classical cost functions in scheduling are earliness/tardiness costs, makespan,
and activity execution or non-execution costs. CP Optimizer generalizes these
classical cost functions and provides a set of basic expressions that can be
combined together; this allows you to express a large spectrum of scheduling cost
functions that can be efficiently exploited by the CP Optimizer search.
For the description of the symbolic notation used throughout this section, see
“Notation” on page 89.
Piecewise linear and stepwise functions
Describes piecewise linear and stepwise functions as related to scheduling.
In CP Optimizer, piecewise linear functions are typically used to model a known
function of time, for instance the cost incurred for completing an activity after a
known date t. Stepwise functions are typically used to model the efficiency of a
resource over time.
Apiecewise linear function F(t) is defined by a tuple F = piecewise(S, T, t0,v
0)
where:
Chapter 2. OPL, the modeling language 69
For a complete description of the OPL syntax of a piecewise linear function, see
piecewise and pwlFunction in the OPL Language Quick Reference.
Astepwise function is a special case of the piecewise linear function, where all
slopes are equal to 0 and the domain and image of Fare integer. A stepwise
function F(t) is defined by a tuple F = stepwise(V, T) where:
For a complete description of the OPL syntax of a stepwise linear function, see
stepwise and stepFunction in the OPL Language Quick Reference.
Examples
The following piecewise and stepwise function are depicted in the diagram, below.
vA V-shape function with value 0 at x= 10, slope −1 before x= 10 and slope s
afterwards:
pwlFunction F1 = piecewise{ -1->10; s } (10, 0);
vAn array of V-shaped functions indexed by iin [1..n] with value 0 at T[i], slope
−U[i] before T[i] and slope V [i] afterwards (T,Uand Vare data integer arrays):
70 OPL Language Reference Manual
pwlFunction F[i in 1..n] = piecewise{ -U[i]->T[i]; V[i] } (T[i],0);
vA stepwise function with value 0 before 0, 100 on [0, 20), value 60 on [20, 30),
and value 100 later on:
stepFunction F2 = stepwise{ 0->0; 100->20; 60->30; 100 };
vA stepwise function with value 0 everywhere except on intervals [7i, 7i+5) for i
in [0, 51] where the value is 100:
stepFunction F3 = stepwise(i in 0..51, p in 0..1) { 100*p -> (7*i)+(5*p)
;0};
Interval variables
Describes a basic building block of scheduling, the interval.
Informally speaking, an interval variable represents an interval of time during
which something happens (a task, an activity is carried out) and whose position in
time is an unknown of the scheduling problem. An interval is characterized by a
start value, an end value and a size. An important feature of interval variables is
the fact that they can be optional; that is, one can decide not to consider them in
the solution schedule. This concept is crucial in applications that present at least
some of the following features:
voptional activities (operations, tasks) that can be left unperformed (with an
impact on the cost); examples include externalized, maintenance or control tasks
vactivities that can execute on a set of alternative resources (machines,
manpower) with possibly different characteristics (speed, calendar) and
compatibility constraints
voperations that can be processed in different temporal modes (for instance in
series or in parallel)
valternative modes for executing a given activity, each mode specifying a
particular combination of resources
valternative processes for executing a given production order, a process being
specified as a sequence of operations requiring resources
vhierarchical description of a project as a work-breakdown structure with tasks
decomposed into sub-tasks, part of the project being optional (with an impact on
the cost if unperformed), and so forth.
Chapter 2. OPL, the modeling language 71
Formally, an interval variable ais a variable whose domain dom(a) is a subset of
. An interval variable is said to be fixed if its
domain is reduced to a singleton; that is, if denotes a fixed interval variable:
vinterval is absent: =;or
vinterval is present: = [s,e)
Absent interval variables have special meaning. Informally speaking, an absent
interval variable is not considered by any constraint or expression on interval
variables it is involved in. For example, if an absent interval variable is used in a
noOverlap constraint, the constraint will behave as if the interval was never
specified to the constraint. If an absent interval variable ais used in a precedence
constraint between interval variables aand bthis constraint does not impact
interval variable b. Each constraint specifies how it handles absent interval
variables.
The semantics of constraints defined over interval variables is described by the
properties that fixed intervals must have in order the constraint to be true. If a
fixed interval is present and such that =[s, e), we will denote s( ) its
integer start value s,e() its integer end value eand l( ) its positive integer
length defined as e( )−s( ). The presence status x( ) will be equal to 1. For a
fixed interval that is absent, x() = 0 and the start, end and length are undefined.
Until a solution is found it may not be known whether an interval will be present
or not. In this case we say that the interval is optional. To be precise, an interval is
said to be absent when dom(a)={ }, present when dom(a) and optional in all
other cases.
Intensity and size
Sometimes the intensity of “work” is not the same during the whole interval. For
example let’s consider a worker who does not work during weekends (his work
intensity during weekends is 0%) and on Friday he works only for half a day (his
intensity during Friday is 50%). For this worker, 7 man-days of work will span for
longer than just 7 days. In this example 7 man-days represents what we call the
size of the interval; that is, what the length of the interval would be if the intensity
function was always at 100%.
To model such situations, you can specify a range for the size of an interval
variable and an integer stepwise intensity function F. For a fixed present interval
the following relation will be enforced at any solution between the start, end,
size sz of the interval and the integer granularity G(by default, the intensity
function is expressed as a percentage so the granularity Gis 100):
That is, the length of the interval will be at least long enough to cover the work
requirements given by the interval size, taking into account the intensity function.
However, any over-estimation is always strictly less than one work unit.
72 OPL Language Reference Manual
If no intensity is specified, it is supposed to be the constant full intensity function
= 100% so in that case sz(a)=l(a). Note that the size is not defined for
absent intervals.
Important:
The intensity step function Fshould be a stepwise function with integer values and
is not allowed to exceed the granularity (100 by default).
The following figure depicts an interval variable of size 14 with its intensity
function. A valid solution is represented where the interval starts at 10 and ends at
27. Indeed in this case:
OPL formulation
Typically, the problem structure will indicate if an interval can be optional or not,
and the keyword optional is used (or not) in the definition of the interval variable.
In the case where the optionality depends on input data, you can specify a boolean
parameter to the optionality field: optional(true) being equivalent to optional and
optional(false) being equivalent to the omission of optional.
A window [StartMin,EndMax] can be specified to restrict the position of the
interval variable. By default, an interval variable will start after 0 and end before
maxint/2. The fixed size or the size range for the interval is specified with the size
keyword. Note that these bounds are taken into account only when the interval
variable is present in the final schedule, that is, they allow specifying conditional
bounds on the interval variable would the interval be present in the final schedule. For
absent intervals, they are just ignored.
dvar interval a [optional[(IsOptional)]]
[in StartMin..EndMax]
[size SZ | in SZMin .. SZMax]
[intensity F]
Where:
int IsOptional, StartMin, EndMax, SZ, SZMin, SZMax;
stepFunction F;
-maxint/2+1<=StartMin <= maxint/2 - 1
Chapter 2. OPL, the modeling language 73
-maxint/2+1<=EndMax <= maxint/2 - 1
0 <= SZ <= maxint/2 - 1
0 <= SZMin <= maxint/2 - 1
0 <= SZMax <= maxint/2 - 1
Examples
For examples of using interval, see the CP keywords interval, optional, size, and
intensity in the OPL Language Quick Reference.
Display of interval variable domain
The domain of an interval variable is displayed as shown in this example:
A1[0..1: 10..990 -- (5..10)5..990 --> 25..1000]
After the name of the interval variable (here A1), the first range (here 0..1)
represents the domain of the boolean presence status of the interval variable. Thus
0..1 represents an optional interval variable whose status has still not been fixed, 0
an absent interval variable and1apresent interval variable.
The remaining fields describe the position of the interval variable, it is omitted if
the interval variable is absent as this information is not relevant in this case. Thus,
an absent interval variable is displayed as:
A1[0]
When the interval variable is possibly present:
- the first range in the remaining fields represents the domain of the interval start
- the second range (between parenthesis) represents the domain of the interval size
- the third range represents the domain of the interval length
- the fourth and last range represents the domain of the interval end
Note that the second range may be omitted in case the size and length of the
interval variable are necessarily equal.
When the values are fixed, the ranges min..max are replaced by a single value. For
instance, the following display represents a fixed interval variable of size 5 that is
present, starts at 10 and ends at 35:
A1[1: 10 -- (5)25 --> 35]
Unary constraints on interval variables
CP Optimizer provides constraints for modeling restrictions that an interval cannot
start, cannot end or cannot overlap a set of fixed dates.
Let denote a fixed interval and Fan integer stepwise function.
vForbidden start. The constraint forbidStart(,F), states that whenever the
interval is present, it cannot start at a value twhere F(t)=0.
vForbidden end. The constraint forbidEnd(,F), states that whenever the interval
is present, it cannot end at a value twhere F(t−1)=0.
74 OPL Language Reference Manual
vForbidden extent. The constraint forbidExtent( , F), states that whenever the
interval is present, it cannot overlap a point twhere F(t)=0.
More formally:
For syntax and examples of these constraints, see forbidEnd, forbidExtent, and
forbidStart in the OPL Language Quick Reference. Note that none of these constraints
can be used in meta-constraints.
Precedence constraints between interval variables
Precedence constraints are common scheduling constraints used to restrict the
relative position of interval variables in a solution.
For example a precedence constraint can model the fact that an activity amust end
before activity bstarts (optionally with some minimum delay z). If one or both or
the interval variables of the precedence constraint is absent, then the precedence is
systematically considered to be true; therefore it does not impact the schedule.
More formally, the semantics of the relation TC( , , z) on a pair of fixed
intervals , and for a value zdepending on the constraint type TC is given in
the following table.
For syntax and examples, see the following functions described in the OPL
Language Quick Reference. Note that none of these constraints may be used in a
meta-constraint.
vendAtEnd
vendAtStart
vendBeforeEnd
vendBeforeStart
vstartAtEnd
vstartAtStart
vstartBeforeEnd
vstartBeforeStart
Chapter 2. OPL, the modeling language 75
Constraints on groups of interval variables
Describes constraints that act to encapsulate a group of intervals together.
The main purpose of a group constraint is to encapsulate a group of interval
variables into one effective higher level interval.
Three “interval grouping” constraints are available: span, alternative, and
synchronize.
Span constraint: The constraint span(a,{b1, .., bn}) states that the interval aspans
over all present intervals from the set {b1, .., bn}. That is, interval astarts together
with the first present interval from {b1, .., bn} and ends together with the last one.
Alternative constraint: The constraint alternative(a,{b1, .., bn}) models an exclusive
alternative between {b1, .., bn}. If interval ais present then exactly one of intervals
{b1, .., bn} is present and astarts and ends together with this chosen one. The
alternative constraint can also be specified by a non-negative integer cardinality c,
alternative(a,{b1, .., bn}, c). In this case, it is not 1 but cinterval variables that will be
selected among the set {b1, .., bn} and those cselected intervals will have to start
and end together with interval variable a.
Synchronize constraint: The constraint synchronize(a,{b1, .., bn}) makes intervals b1.
..bnstart and end together with interval a(if ais present).
Note that the alternative, span, and synchronize constraints cannot be used in
meta-constraints.
For syntax and examples, see the functions as described in the OPL Language Quick
Reference.
valternative
vspan
vsynchronize
A logical constraint between interval variables: presenceOf
The presence constraint states that a certain interval must be present in the
solution.
76 OPL Language Reference Manual
The semantics of the presence constraint on a fixed interval is simply:
presenceOf( )↔x( ). The truth value of this constraint can be used in arithmetical
expressions, and thereby restricted by logical constraints.
This constraint can be used in meta-constraints to indicate, for example, that there
may be two optional intervals aand b; if interval ais present then bmust be
present as well. This is modelled by the constraint presenceOf(a)→presenceOf(b).
OPL formulation
The constraint to express that the interval variable must be present:
presenceOf(a);
where:
dvar interval a;
For an example of the presenceOf constraint see presenceOf in the OPL Language
Quick Reference.
Expressions on interval variables
Integer and numerical expressions are available to access or evaluate different
attributes of an interval variable.
These expressions can be used, for example, to define a term for the cost function
or to connect interval variables to integer and floating point expressions.
The integer expressions are startOf,endOf,lengthOf, and sizeOf and they provide
access to the different attributes of an interval variable. Special care must be taken
for optional intervals, as an integer value dval must be specified which represents
the value of the expression when the interval is absent. If this value is omitted, it is
supposed to be 0. For the syntax of integer expressions, see endOf, lengthOf,
sizeOf, and startOf in the OPL Language Quick Reference.
The numerical expressions are startEval,endEval,lengthEval, and sizeEval, and they
allow evaluation of a piecewise linear function on a given bound of an interval. As
with integer expressions, in the case of optional intervals an integer value dval
must be specified which represents the value of the expression when the interval is
absent. If this value is omitted, it is supposed to be 0. For the syntax and examples
of the use of a numerical expression, see endEval, lengthEval, sizeEval, and
startEval in the OPL Language Quick Reference.
Let denote a fixed interval variable. The semantics of these expressions is shown
in the table.
Chapter 2. OPL, the modeling language 77
Sequencing of interval variables
Describes a basic building block of scheduling, the interval sequence.
⊂π∀
An interval sequence variable is defined on a set of interval variables A. Informally
speaking, the value of an interval sequence variable represents a total ordering of
the interval variables of A. Note that any absent interval variables are not
considered in the ordering.
More formally, an interval sequence variable pon a set interval variables A
represents a decision variable whose possible values are all the permutations of the
intervals of A. Let be a set of fixed intervals and ndenote the cardinality of .
Note that the sequence alone does not enforce any constraint on the relative
position of intervals end-points. For instance, an interval variable acould be
sequenced before an interval variable bin a sequence pwithout any impact on the
relative position between the start/end points of aand b(acould still be fixed to
start after the end of b). This is because different semantics can be used to define
78 OPL Language Reference Manual
how a sequence constrains the positions of intervals. We will see later how the
noOverlap constraint implements one of these possible semantics.
The sequence variable also allows associating a fixed non-negative integer type
with each interval variable in the sequence. In particular, these integers are used by
the noOverlap constraint. T(p,a) denotes the fixed non-negative integer type of
interval variable ain the sequence variable p.
Constraints on sequence variables
The following constraints are available on sequence variables:
vfirst (scheduling)(p,a) states that if interval ais present, then it will be the first
interval of the sequence p.
vlast (scheduling)(p,a) states that if interval ais present, then it will be the last
interval of the sequence p.
vbefore(p,a,b) states that if both intervals aand bare present then awill appear
before bin the sequence p.
vprev (scheduling)(p,a,b) states that if both intervals aand bare present then a
will be just before bin the sequence p, that is, it will appear before band no
other interval will be sequenced between aand bin the sequence.
The formal semantics of these basic constraints is shown in the following table.
The no overlap constraint
The no overlap constraint on an interval sequence variable pstates that the sequence
defines a chain of non-overlapping intervals, any interval in the chain being
constrained to end before the start of the next interval in the chain. This constraint
is typically useful for modelling disjunctive resources.
More formally, the condition for a permutation value for sequence pto satisfy
the noOverlap constraints is defined as:
If a transition distance matrix Mis specified, it defines the minimal non-negative
distance that must separate two consecutive intervals in the sequence.
More formally, if T(p,a) denotes the non-negative integer type of interval ain the
sequence variable p:
Chapter 2. OPL, the modeling language 79
A sequence variable together with a no-overlap constraint using it are illustrated in
this figure:
The sameSequence and sameCommonSubsequence constraints
The sameSequence and sameCommonSubsequence constraints are binary constraints on
a pair of sequence variables p1and p2. These constraints state that the relative
position of some related subsets of interval variables is the same in both sequences.
Typically, if ais before bin sequence p1then the interval related to ais before the
interval related to bin sequence p2. Here are two examples of use-cases for these
constraints.
vFirst in/first out and no-bypass constraints. In some physical systems, like trains
on a single line railway, or items on a conveyor belt, bypassing is not possible
and items must enter and exit a given section of the system in the same order. If
entering and exiting the sections or the junctions of the system is modeled by
two related interval variables, those constraints can be modeled by sameSequence
constraints (or sameCommonSubsequence constraints if the items do not follow the
same path) on the sequences of entering and exiting intervals. A classic example
of such a constraint is the permutation flow-shop scheduling problem.
vScenario-based approaches for scheduling with uncertainties. In presence of
uncertainties (for instance, uncertain activity durations) one may be interested in
building sequences of activities on unary resources that optimize some
robustness or statistical criterion (for instance the expected makespan). A
scenario is a sub-model that defines a particular realization of the uncertainties
in the environment. As one must find robust sequences that optimize some
criterion over all scenarios, the different sequences of a given unary resource
across all scenarios are linked with sameSequence constraints.
A constraint sameCommonSubsequence(p1,p
2,B
1,B
2)is defined on two sequence
variables p1(defined on a set of interval variables A1) and p2(defined on a set of
interval variables A2), with two arrays of interval variables B1and B2of identical
size and such that B1⊂A1,B
2⊂A2. The constraint states that the sub-sequences
80 OPL Language Reference Manual
defined by p1and p2by only considering the pairs of present intervals (B1[i], B2[i])
are identical modulo the mapping between intervals B1[i] and B2[i].
For instance, let's suppose sequence variable p1defined on interval variables
{a,b,c,d,e,f} and sequence variable p2defined on interval variables {u,v,w,x,y,z} with a
constraint sameCommonSubsequence(p1,p
2, [a,c,d,e,f], [u,w,v,x,y]). Assuming that, in a
solution, interval variables dand yare absent while all the other interval variables
are present. The pair of permutations π1=(c,f,a,e,b) and π2=(w,v,u,x) satisfies the
sameCommonSubsequence constraints. Indeed if one considers only pairs of present
intervals, the mapping B1,B2maps [a,c,e] with [u,w,x]. Sub-sequence of π1over
{a,c,e} is (c,a,e), sub-sequence of π2over {u,w,x} is (w,u,x) and one can see that the
two sub-sequences are the same modulo the mapping.
More formally, let ndenote the size of both interval variable arrays B1and B2. The
condition for permutations π1and π2of sequencing variables p1and p2to satisfy the
constraint sameCommonSubsequence(π1,π2,B
1,B
2)is defined as:
∀i, j ∈[1, n], i ≠j, x(B1[i]), x(B1[j]), x(B2[i]), x(B2[j]), (π1(B1[i]) < π1(B1[j]) ⇔π2(B2[i]) <
π2(B2[j])
A constraint sameSequence(p1,p
2,B
1,B
2)is defined on two sequence variables of
identical size p1(defined on a set of interval variables A1) and p2(defined on a set
of interval variables A2) with two arrays of interval variables B1and B2that are an
ordering of sets A1and A2. The constraint states that sequences p1and p2are
identical modulo a mapping between intervals B1[i] and B2[i].AsameSequence
constraint is stronger than a sameCommonSubsequence constraint; it additionally
imposes that arrays B1and B2contain all the definition intervals of their relative
sequence variables and that the presence status of mapped interval variables B1[i]
and B2[i] is the same.
More formally, let ndenote the size of both sequence variables. The condition for
permutations π1and π2of the sequence variables p1and p2to satisfy constraint
sameSequence(π1,π2,B
1,B
2)is defined as:
∀i∈[1, n], π1(B1[i]) = π2(B2[i])
When parameters B1and B2are omitted in the definition of the above constraints,
we assume B1and B2are the definition intervals of p1and p2, following the
definition order of interval variables in the sequence.
Syntax and examples
For syntax and examples of use of the sequence interval variable, see sequence in
the OPL Language Quick Reference.
For the syntax and examples of use of the no overlap constraint, which needs to be
defined as a set of integer triples, see noOverlap in the OPL Language Quick
Reference.
For the syntax and examples of the other constraints available on an interval
sequence variable, see first, last (scheduling), prev (scheduling), before,
sameSequence, and sameCommonSubsequence in the OPL Language Quick
Reference. (Note that there are similarly-named constraints available for set
operations in OPL.)
Chapter 2. OPL, the modeling language 81
Note that none of the constraints mentioned in this section can be used in a
meta-constraint.
Sequence variables can be viewed in a Gantt chart in the CPLEX Studio IDE. A
simple example is given in Viewing the results of scheduling problems.
Cumulative functions
Describes the cumulative function.
In scheduling problems involving cumulative resources (also known as renewable
resources), the cumulated usage of the resource by the activities is usually
represented by a function of time. An activity usually increases the cumulated
resource usage function at its start time and decreases it when it releases the
resource at its end time (pulse function). For resources that can be produced and
consumed by activities (for instance the content of an inventory or a tank), the
resource level can also be described as a function of time; production activities will
increase the resource level whereas consuming activities will decrease it. In these
type of problems, the cumulated contribution of activities on the resource can be
represented by a function of time and constraints can be modeled on this function,
for instance a maximal or a safety level.
CP Optimizer introduces the notion of the cumulative function expression, which is a
function that represents the sum of individual contributions of intervals. A panel of
elementary cumul function expressions is available to describe the individual
contribution of an interval variable (or a fixed interval of time) which cover the
main use-cases mentioned above: pulse for usage of a cumulative resource, step for
resource production/consumption. When the elementary cumulative function
expressions that define a cumul function expression are fixed (and thus, so are
their related intervals), the expression is fixed. CP Optimizer provides several
constraints over cumul function expressions. These constraints allow restricting the
possible values of the function over the complete horizon or over some fixed or
variable interval. For applications where the actual quantity of resource that is
used, produced or consumed by intervals is an unknown of the problem,
expressions are available for constraining these quantities.
Cumul function expressions
Let denote the set of all functions from . A cumul function expression
fis an expression whose value is a function of and thus, whose domain dom(f)
is a subset of . Let and and abe an interval
variable, we consider the following elementary cumul function expressions
illustrated in the following figure: pulse(u, v, h), step(u, h), pulse(a, h), pulse(a, hmin,
hmax), stepAtStart(a, h), stepAtStart(a, hmin,hmax), stepAtEnd(a, h), and stepAtEnd(a,
hmin,hmax).
82 OPL Language Reference Manual
More formally, let and and we define the following particular
functions of
The semantics of the elementary function expressions is listed in the following
table, together with the formal definition of their domain. The function set is
equal to the singleton if ; that is, if interval variable ais
possibly absent, and equal to the empty set otherwise.
Chapter 2. OPL, the modeling language 83
Constraints on cumul function expressions
The following constraints can be expressed on a cumul function expression f. Let
and and abe an interval variable:
valwaysIn(f, u, v, hmin,hmax) means that the values of function fmust remain in the
range [hmin,hmax] everywhere on the interval [u, v).
valwaysIn(f, a, hmin,hmax) means that if interval ais present, the values of function f
must remain in the range [hmin,hmax] between the start and the end of interval
variable a.
vf≤hmax means that function fcannot take values greater than hmax.Itis
semantically equivalent to .
vf≥hmin means that function fcannot take values lower than hmin.Itis
semantically equivalent to
More formally:
Expressions on cumulative functions
The following elementary cumul function expressions are based on an interval
variable a:pulse(a,h), pulse(a,hmin,hmax), stepAtStart(a,h), stepAtStart(a,hmin,hmax),
stepAtEnd(a,h), and stepAtEnd(a,hmin,hmax).
Some of these expressions define a range [hmin,hmax] of possible values for the
actual height of the function when the interval variable ais present. The actual
height is an unknown of the problem. CP Optimizer provides some integer
expressions to control this height. These expressions are based on the notion of the
contribution of a given interval variable ato the (possibly composite) cumul
function expression f. This contribution is defined as the sum of all the elementary
84 OPL Language Reference Manual
cumul function expressions based on ain f. This contribution is a discrete function
that can change value only at the start and at the end of interval aand is equal to
0 before the start of a.
For instance, let aand bbe two interval variables and a cumul function expression
fdefined by: f=pulse(a,3)+pulse(a,2)−stepAtEnd(a,1)+stepAtStart(b,2)−
stepAtEnd(b, 3). The contribution of ato fis the function pulse(a,3)+pulse(a,2)−
stepAtEnd(a, 1) and the contribution of bto fis the function stepAtStart(b,2)−
stepAtEnd(b, 3).
If interval ais present, the expression heightAtStart(a,f) returns the value of the
contribution of ato fevaluated at the start of athat is, it measures the contribution
of interval ato cumul function expression fat its start point. Similarly, the
expression heightAtEnd(a,f) returns the value of the contribution of ato fevaluated
at the end of athat is, it measures the contribution of interval ato cumul function
expression fat its end point. An additional integer value dval can be specified at
the construction of the expression, which will be the value returned by the
expression when the interval is absent. Oherwise, if no value is specified, the
expression will be equal to 0 when the interval is absent.
In the example above, assuming both interval aand bto be present we would get:
heightAtStart(a,f)=5,heightAtEnd(a,f)=4,heightAtStart(b,f)=2,heightAtEnd(b,f)=
−1. Of course, in general when using ranges for the height of elementary cumul
function expressions, the heightAtStart/End expressions will not be fixed until all
the heights have been fixed by the search.
Syntax and examples
For the syntax and examples of use of a cumulative function see cumulFunction,
pulse, step, stepAtEnd, and stepAtStart in the OPL Language Quick Reference.
The results of a solved scheduling model can be viewed in the CPLEX Studio IDE.
An example of results containing a cumulative function is provided in Viewing the
results of cumulative functions in the IDE.
These are the constraints available on cumulative function expressions:
vf <= hmax;
vhmin <= f;
valwaysIn(f, u, v, hmin, hmax)
valwaysIn(f, a, hmin, hmax)
Note that these constraints cannot be used in meta-constraints.
Chapter 2. OPL, the modeling language 85
The following expressions are available on cumulative functions:
vdexpr int h = heightAtStart(a,f[,dval]);
vdexpr int h =heightAtEnd(a,f[,dval]);
More information on these constraints and expressions is available in the OPL
Language Quick Reference.
State functions
Describes the state function.
Some scheduling problems involve reasoning with resources whose state may
change over time. The state of the resource can change because of the scheduled
activities or because of exogenous events; yet some activities in the schedule may
need a particular condition on the resource state to be true in order to execute. For
instance, the temperature of an oven may change due to an activity that sets the
oven temperature to a value v, and a cooking activity may follow that requires the
oven temperature to start at and maintain a temperature level v' throughout its
execution. Furthermore, the transition between two states is not always
instantaneous and a transition time may be needed for the resource to switch from
a state vto a state v'.
CP Optimizer introduces the notion of state function which is used to describe the
evolution of a given feature of the environment. The possible evolution of this
feature is constrained by interval variables of the problem. The main difference
between state functions and cumulative functions is that interval variables have an
incremental effect on cumul functions (increasing or decreasing the function value)
whereas they have an absolute effect on state functions (requiring the function
value to be equal to a particular state or in a set of possible states).
Informally speaking, a state function is a set of non-overlapping intervals over
which the function maintains a particular non-negative integer state. In between
those intervals, the state of the function is not defined, typically because of an
ongoing transition between two states. For instance for an oven with three possible
temperature levels identified by indexes 0, 1 and 2 we could have:
v[start=0, end=100): state=0,
v[start=150, end=250): state=1,
v[start=250, end=300): state=1,
v[start=320, end=420): state=2,
v[start=460, end=560): state=0, ...
Constraints are available to restrict the evolution of a state function. These
constraints allow you to specify:
vThat the state of the function must be defined and should remain equal to a
given non-negative state everywhere over a given fixed or variable interval
(alwaysEqual).
vThat the state of the function must be defined and should remain constant (no
matter its value) everywhere over a given fixed or variable interval
(alwaysConstant).
vThat intervals requiring the state of the function to be defined cannot overlap a
given fixed or variable interval (alwaysNoState).
vThat everywhere over a given fixed or variable interval, the state of the function,
if defined, must remain within a given range of non-negative states [vmin,vmax]
(alwaysIn).
86 OPL Language Reference Manual
Additionally, the two first constraints can be complemented to specify that the
given fixed or variable interval should have its start and/or end point
synchronized with the start and/or end point of the interval of the state function
that maintains the required state. This is the notion of start and end alignment
which is particularly useful for modelling parallel batches. For instance in the oven
example above, all interval variables that would require an oven temperature of
level 1 and specify a start and end alignment, if executed over the interval [150,
250) would have to start exactly at 150 and end at 250. This is depicted in the
following figure where a1and a2are two start and end aligned interval variables, a3
is start aligned only and a4is not aligned at all.
State functions and transition distance
A state function fis a decision variable whose value is a set of non-overlapping
intervals, each interval [si,ei) being associated a non-negative integer value vithat
represents the state of the function over the interval.
For instance, in the example of the oven introduced previously, we would have
(200) = 1, s( , 200) = 150 and e(f, 200) = 250.
A state function can be associated with a transition distance. The transition distance
defines the minimal distance that must separate two consecutive states in the state
function. More formally, if M[v,v'] represents a transition distance matrix between
state vand state v', the transition distance means that:
Chapter 2. OPL, the modeling language 87
The transition distance matrix Mmust satisfy the triangular inequality. For
instance, in the example of the oven, the state function depicted in the previous
figure is consistent with the following transition distance:
Constraints on state functions
If fis a state function, aan interval variable, v,vmin,vmax non-negative integers and
algns,algnetwo boolean values:
vThe constraint alwaysEqual(f,a,v,algns,algne) specifies that whenever ais
present, state function fmust be defined everywhere between the start and the
end of interval aand be constant and equal to non-negative value vover this
interval. If algnsis true, it means that interval a is start-aligned with f: Interval a
must start at the beginning of the interval where fis maintained in state s.If
algneis true, it means that interval ais end-aligned with f: Interval amust end at
the end of the interval where fis maintained in state s. More formally:
vThe constraint alwaysConstant(f, a,algns,algne) specifies that whenever ais
present, state function fmust be defined everywhere between the start and the
end of interval aand be constant over this interval. More formally:
vThe constraint alwaysNoState(f,a) specifies that whenever ais present, state
function fcannot provide any valid state between the start and the end of
interval a. As a consequence, any interval constrained with alwaysEqual or
alwaysConstant on this function and thus requiring the function to be defined
cannot overlap interval a. Formally:
vThe constraint alwaysIn(f,a,vmin,vmax) specifies that whenever ais present,
everywhere between the start and the end of interval athe state of function f,if
defined, must belong to the range [vmin,vmax] where 0 ≤Vmin ≤Vmax. Formally:
Syntax and examples
For the syntax and examples of use of a state function see stateFunction.
The results of a solved scheduling model can be viewed in the CPLEX Studio IDE.
An example of results containing a state function is provided in Viewing the
results of state functions in a Gantt chart.
88 OPL Language Reference Manual
Another example featuring a transition and the alwaysEqual constraint is shown
below.
The following list includes the constraints available on a state function. A full
description and example for each syntax is available in the OPL Language Quick
Reference.
valwaysEqual(f,s,e,v[,aligns,aligne]);
valwaysEqual(f,a,v[,aligns,aligne]);
valwaysConstant(f,s,e,[,aligns,aligne]);
valwaysConstant(f,a,[,aligns,aligne]);
valwaysNoState(f,s,e);
valwaysNoState(f,a);
valwaysIn(f,u,v,hmin,hmax);
valwaysIn(f,a,hmin,hmax);
Note that these constraints cannot be used in meta-constraints.
Example with stateFunction, transition, and alwaysEqual.
A machine can be equipped with a tool among a set of npossible tools. Each
operation oexecuted on the machine needs a specific tool RequiredTool[o]. The
machine can process several operation simultaneously provided these operations
are compatible with the tool currently installed on the machine. Changing the tool
installed on the machine needs some constant set-up time which is supposed to be
independent from the tools.
int nbTools = ...;
int nbOps = ...;
int setupTime = ...;
range Tools = 1..nbTools;
range Operations = 1..nbOps;
int Duration [Operations] = ...;
int RequiredTool [Operations] = ...;
dvar interval op[o in Operations] size Duration[o];
tuple triplet { int tl1; int tl2; int value; };
{ triplet } Transition = { <tl1,tl2,setupTIme> } tl1, tl2 in Tools };
stateFunction machineTool with Transition;
constraints {
forall(o in Operations) {
alwaysEqual(machineTool, op[o], RequiredTool[o]);
}
}
Notation
The main types of notation used throughout the scheduling section are defined
here.
Vectors are denoted by capital letters, for example Y. The size of a vector Yis
denoted |Y|. If n=|Y|, the vector is denoted Y=(y1,...,yn).
Chapter 2. OPL, the modeling language 89
90 OPL Language Reference Manual
Chapter 3. IBM ILOG Script for OPL
Describes the structure and built-in values and functions of the scripting language.
Language structure
Presents the structure of the IBM ILOG Script language for OPL: the language
constructs, the elements from which expressions can be constructed, and the
possible types of statement.
Syntax
What composes a scripting statement, compound statements, comments, identifiers.
General
Provides a general overview of OPL syntax.
A script comprises a sequence of statements. An expression can also be used
whenever a statement is expected, in which case its value is ignored and only its
side effect is taken into account.
You can put multiple statements or expressions on a single line if you separate
them with a semi-colon (;), for example, the following two scripts are equivalent:
Script1
writeln("Hello, world")
x = x+1
if (x > 10) writeln("Too big")
Script2
writeln("Hello, world"); x = x+1; if (x > 10) writeln("Too big")
Compound statements
Explains the use of compound statements in OPL syntax.
A compound statement is a sequence of statements and expressions enclosed in
curly brackets ({}). It can be used to perform multiple tasks whenever a single
statement is expected, for example, in the following conditional statement, the
three statements and expressions in curly brackets are executed when the condition
a>bis true:
if(a>b){
varc=a
a=b
b=c
}
The last statement or expression before a closing curly bracket does not need to be
followed by a semicolon, even if it is on the same line. For example, the following
program is syntactically correct and is equivalent to the previous one:
if(a>b){varc=a;a=b;b=c}
© Copyright IBM Corp. 1987, 2015 91
Comments
Explains the syntax of comments in OPL.
Script supports two different styles of comments:
vSingle line comments: A single line comment starts with // and stops at the
end of the line.
Example:
x = x+1 // Increment x,
y = y-1 // then decrement y.
vMultiple line comments: To span on more than one line, comments must start
with a /* and ends with a */; Nested multiple line comments are not allowed.
Example:
/* The following statement
increments x. */
x = x+1
/* The following statement
decrements y. */
y=y/*Acomment can be inserted here */ -1
Identifiers
Shows how to use identifiers in OPL syntax.
Identifiers are used to name script variables and functions. An identifier starts with
either a letter or an underscore, and is followed by a sequence of letters, digits,
and underscores.
These are examples of identifiers:
car
x12
main_window
_foo
The language is case-sensitive, so that the uppercase letters A-Z are distinct from
the lowercase letters a-z. For example, the identifiers car and Car are distinct.
Expressions in IBM ILOG Script
Expressions are a combination of literals, script variables, special keywords, and
operators.
Literals
Explains the use of literals in IBM ILOG Script.
Literals can represent the following elements:
vNumbers, for example: 12 14.5 1.7e-100
vStrings, for example, "Ford" "Hello world\n"
vBooleans, either true or false
vThe null value: null.
For further details about number and string literal syntax, see “Numbers” on page
103 and “IBM ILOG Script strings” on page 110.
Operators
Explains the use of and precedence of operators in IBM ILOG Script.
92 OPL Language Reference Manual
The precedence of operators determines the order in which they are applied when
an expression is evaluated. You can override operator precedence using
parentheses.
For a list of IBM ILOG Script operators, see Operators.
Syntax of different types of expression
Describes the syntax of several types of expression in OPL.
The following types of expression are described:
vA reference to a script variable
vAccess to a property
vAssignment operators
vFunction calls
vSpecial keywords
vSpecial operators
vOther operators
Note:
For C/C++ programmers: The syntax of Script expressions is very close to the C
and C++ syntax. Expressions include assignments, function calls, property access,
and so on.
Script variable reference
Provides a reference for the syntax of script variables in OPL.
Table 7. Reference syntax of script variables
Syntax Effect
variable Returns the value of variable. See
“Identifiers” on page 92 for the syntax of
script variables. If variable does not exist,
an error is signalled. This is not the same as
referencing an existing script variable whose
value is the undefined value, which is legal
and returns the undefined value. When used
in the body of a with statement, a variable
reference is first looked up as a property of
the current default value.
Property access
Provides a reference for the syntax used to access properties in OPL.
There are two ways of accessing a property value.
Chapter 3. IBM ILOG Script for OPL 93
Table 8. Property access syntax
Syntax Effect
value.name Returns the value of the name property of
value, or the undefined value if this
property is not defined. See “Identifiers” on
page 92 for the syntax of name.
Examples:str.length getCar().name
Because name must be a valid identifier, this
form cannot be used to access properties
which do not have a valid identifier syntax.
For example, the numeric properties of an
array cannot be accessed this
way:myArray.10 // Illegal syntax
For these properties, use the second syntax.
value[name] Same as the previous syntax, except that this
time name is an evaluated expression which
gives the property name.
Examples:str["length"] // Same as str.length
getCar()[getPropertyName()] myArray[10]
myArray[i+1]
Assignment operators
Provides a reference for the syntax of assignment operators in OPL.
The equals (=) operator can be used to assign a new value to a script variable or a
property.
Table 9. Assignment operator syntax
Syntax Effect
variable = expression In scripting, all objects are assigned by
reference, except strings, numbers and
Booleans, which are assigned by value. See
the ECMA standard for details.
Example:x = y+1
The whole expression returns the value of
expression.
value.name = expression
value[name] = expression
Assigns the value of expression to the given
property.
If value does not have such a property, then
if it is either an array or an object, the
property is created; otherwise, an error is
signalled.
Examplecar.name = "Ford" myArray[i] =
myArray[i]+1
The whole expression returns the value of
expression.
In addition, shorthand operators are also defined.
94 OPL Language Reference Manual
Function calls
Provides a reference for the shorthand syntax used with function calls in OPL.
Table 10. Syntactic shorthand
Syntax Shorthand for
++X X = X+1
X++ Same as ++X, but returns the initial value of
Xinstead of its new value.
--X X = X-1
X-- Same as --X, but returns the initial value of
Xinstead of its new value.
X+=Y X=X+Y
X-=Y X=X-Y
X*=Y X=X*Y
X/=Y X=X/Y
X%=Y X=X%Y
X<<=Y X=X<<Y
X>>=Y X=X>>Y
X >>>= Y X = X >>> Y
X&=Y X=X&Y
X^=Y X=X^Y
X|=Y X=X|Y
Table 11. Function call syntax
Syntax Effect
function (arg1, ..., argn) Calls function with the given arguments,
and returns the result of the call.
Examples:parseInt(field) writeln("Hello ",
name) doAction() str.substring(start,
start+length)
The function is typically either a script
variable reference or a property access, but it
can be any expression; the expression must
yield a function value, or an error is
signalled.
Examples:callbacks[i](arg) // Calls the
function in callbacks[i] "foo"() // Error: a
string is not a function
Chapter 3. IBM ILOG Script for OPL 95
Special keywords
Lists special reserved keywords in OPL.
Table 12. Special keyword syntax
Syntax Effect
this When referenced in a method, returns the
current calling object; when referenced in a
constructor, returns the object currently
being initialized. Otherwise, returns the
global object. See “this as a keyword” on
page 124 for examples.
arguments Returns an array containing the arguments
of the current function. When used outside a
function, an error is signalled.
For example, the following function returns
the sum of all its arguments:function sum() {
var res = 0 for (var i=0; i<arguments.length;
i++) res = res+arguments[i] return res }
The call sum(1, 3, 5) returns 9.
Special operators
Lists the special operators in OPL and their syntax.
Table 13. Special operator syntax
Syntax Effect
new constructor(arg1, ..., argn) Calls the constructor with the given
arguments, and returns the created value.
Examples:new Array() new MyCar("Ford",
1975)
The constructor is typically a script variable
reference, but it can be any expression.
Example:new ctors[i](arg) // Invokes
constructor ctors[i]
typeof value Returns a string representing the type of
value, as follows:Array "object" Boolean
"boolean" Date "date" Function "function"
Null "object" Number "number" Object
"object" String "string" Undefined
"undefined"
96 OPL Language Reference Manual
Table 13. Special operator syntax (continued)
Syntax Effect
delete variable Deletes the global script variable variable.
This does not mean that the value in
variable is deleted, but that it is removed
from the global environment.
Example:myVar = "Hello, world" // Create
the global variable myVar delete myVar
writeln(myVar) // Signals an error because
myVar is undefined
If variable is a local variable, an error is
signalled; if variable is not a known
variable, nothing happens.
The whole expression returns true.
For C/C++ programmers: The delete
operator has a radically different meaning in
IBM ILOG Script; in C++, it is used to
delete objects, not script variables and
properties.
delete value.name
delete value[name]
Remove the property name from the object
value.
If value does not contain the name property,
this expression does nothing. If the property
does exist but cannot be deleted, an error is
signalled. If value is not an object, an error
is signalled.
The whole expression returns the true value.
expression1 , expression2 Evaluates expression1 and expression2
sequentially, and returns the value of
expression2. The value of expression1 is
ignored.
The most common use for this operator is
inside for loops, where it can be used to
evaluate several expressions where a single
expression is expected:
for (var i=0, j=0; i<10; i++, j+=2)
{ writeln(j, " is twice as big as ", i); }
Other operators
Provides a reference of other operators used in OPL.
Other operators are described in detail in the section dedicated to the data type
they operate on.
Chapter 3. IBM ILOG Script for OPL 97
Table 14. Other operator syntax
Syntax Effect
-X
X+Y
X-Y
X*Y
X/Y
X%Y
Arithmetic operators.
These operators perform the usual
arithmetic operations. In addition, the +
operator can be used to concatenate strings.
See “Numeric operators” on page 108 and
“String operators” on page 116.
X==Y
X!=Y
Equality operators.
These operators can be used to compare
numbers and strings; see “Numeric
operators” on page 108 and “String
operators” on page 116.
For other types of values, such as dates,
arrays, and objects, the == operator is true
if, and only if, X and Y are the exact same
value. For example:
new Array(10) == new Array(10) false var a
= new Array(10); a == a true
X>Y
X>=Y
X<Y
X<=Y
Relational operators.
These operators can be used to compare
numbers and strings. See “Numeric
operators” on page 108 and “String
operators” on page 116.
~X
X&Y
X|Y
X^Y
X<<Y
X>>Y
X >>> Y
Bitwise operators.
See “Numeric operators” on page 108.
!X
X||Y
X&&Y
condition ?X:Y
Logical operators.
See “Logical operators” on page 118.
Statements
A statement can be a conditional statement, a loop statement, a local script variable
declaration, a function definition, or a default value.
98 OPL Language Reference Manual
Conditional statement
Provides a reference for the syntax of conditional statements in OPL.
Loops
Table 15. Conditional statement syntax
Syntax Effect
if (expression) statement1
[else statement2]
Evaluate expression ; if it is true, execute
statement1 ; otherwise, if statement2 is
provided, execute statement2.
If expression gives a non-Boolean value, this
value is converted to a Boolean value.
Examples:
if (a == b) writeln("They are equal") else
writeln("They are not equal") if
(s.indexOf("a") < 0) { write("The string ", s)
writeln(" doesn’t contains the letter a") }
Table 16. Loop syntax
Syntax Effect
while (expression)
statement
Execute statement repeatedly as long as
expression is true. The test takes place
before each execution of statement.
If expression gives a non-Boolean value, this
value is converted to a Boolean value.
Examples:
while (a*a < b) a = a+1 while (s.length){r=
s.charAt(0)+r s = s.substring(1) }
for ( [ initialize ] ;
[ condition ] ;
[ update ] )
statement
where condition and update are expressions,
and initialize is either an expression or has
the form:
var variable = expression
Evaluate initialize once, if present. Its
value is ignored. If it has the form:var
variable = expression
then variable is declared as a local script
variable and initialized as in the var
statement.
Then, execute statement repeatedly as long
as condition is true. If condition is omitted,
it is taken to be true, which results in an
infinite loop. If condition gives a
non-Boolean value, this value is converted to
a Boolean value.
If present, update is evaluated at each pass
through the loop, after statement and before
condition. Its value is ignored.
Example:for (var i=0; i < a.length; i++) { sum
= sum+a[i] prod = prod*a[i] }
Chapter 3. IBM ILOG Script for OPL 99
Table 16. Loop syntax (continued)
Syntax Effect
for ([var ] variable in expression)
statement
Iterate over the properties of the value of
expression : for each property, variable is
set to a string representing this property, and
statement is executed once.
If the var keyword is present, variable is
declared as a local script variable, as with
the var statement.
For example, the following function takes an
arbitrary value and displays all its properties
and their values:function printProperties(v) {
for (var p in v) writeln(p, " -> ", v[p]) }
Properties listed by the for..in statement
include method properties, which are merely
regular properties whose value is a function
value. For example, the call
printProperties("foo") would
display:length -> 3 toString -> [primitive
method toString] substring -> [primitive
method substring] charAt -> [primitive
method charAt] etc
The only properties which are not listed by
for..in loops are the numeric properties of
arrays.
break Exit the current while,for or for..i n loop,
and continue the execution at the statement
immediately following the loop. This
statement cannot be used outside a loop.
Example:while (i < a.length) { if (a[i] ==
"foo") { foundFoo = true break}i=i+1}//
Execution continues here
continue Stop the current iteration of the current
while,for or for..i n loop, and continue
the execution of the loop with the next
iteration. This statement cannot be used
outside a loop.
Example: for (var i=0; i < a.length; i++) { if
(a[i] < 0) continue writeln("A positive
number: ", a[i]) }
100 OPL Language Reference Manual
Declaration of script variables
Provides a reference for the syntax of script variable declarations in OPL.
Table 17. Declaration syntax for script variables
Syntax Effect
var decl1, ..., decln
where each decli
has the form
variable [ = expression ]
Declares each script variable as a local
variable. If an expression is provided, it is
evaluated and its value is assigned to the
variable as its initial value. Otherwise, the
variable is set to the undefined value.
Examples:var x var name = "Joe" var
average = (a+b)/2, sum, message="Hello"
Inside a function definition
Script variables declared with var are local to the function, and they hide any
global variables with the same names; they have the same status as function
arguments.
For example, in the following program, the script variables sum and res are local to
the average function, as well as the arguments aand b; when average is called,
the global variables with the same names, if any, are temporarily hidden until exit
from the function:
function average(a, b) {
var sum = a+b
var res = sum/2
return res
}
Script variables declared with var at any place in a function body have a scope
which is the entire function body. This is different from local variable scope in C or
C++. For example, in the following function, the variable res declared in the first
branch of the if statement is used in the other branch and in the return
statement:
function max(x, y) {
if(x>y){
var res = x
} else {
res=y
}
return res
}
Outside a function definition
At the same level as function definitions, script variables declared with var are
local to the current program unit. A program unit is a group of statements which is
considered a whole; the exact definition of a program unit depends on the
application in which the script is embedded. Typically, a script file loaded by the
application is treated as a program unit. In this case, variables declared with var at
the file top level are local to this file, and they hide any global variables with the
same names.
For example, suppose that a file contains the following program:
Chapter 3. IBM ILOG Script for OPL 101
var count = 0
function NextNumber() {
count = count+1
return count
}
When this file is loaded, the function NextNumber becomes visible to the whole
application, while count remains local to the loaded program unit and is visible
only inside it.
It is an error to declare the same local variable twice in the same scope. For
example, the following program is incorrect because res is declared twice:
function max(x, y) {
if(x>y){
var res = x
} else {
varres=y//Error }
return res
}
Function definitions
Provides a reference for the syntax of function definitions in OPL.
Table 18. Function definition syntax
Syntax Effect
[static ]
function name(v1, ..., vn)
{statements}
Defines a function name with the given
parameters and body. A function definition
can only take place at the top level; function
definitions cannot be nested.
When the function is called, the script
variables v1, ..., vn are set to the
corresponding argument values, then the
statements are executed. If a return
statement is reached, the function returns
the specified value; otherwise, after the
statements are executed, the function returns
the undefined value.
The number of actual arguments does not
need to match the number of parameters: if
there are fewer arguments than parameters,
the remaining parameters are set to the
undefined value; if there are more
arguments than parameters, the excess
arguments are ignored.
Independently of the parameter mechanism,
the function arguments can be retrieved
using the arguments keyword described in
Table 12 on page 96.
return [ expression ] Returns the value of expression from the
current function. If expression is omitted,
returns the undefined value. The return
statement can only be used in the body of a
function.
102 OPL Language Reference Manual
Defining a function name is operationally the same as assigning a specific function
value to the variable name; thus a function definition is equivalent to:
var name = some function value
The function value can be retrieved from the script variable and manipulated like
any other type of value. For example, the following program defines a function add
and assigns its value to the variable sum, which makes add and sum synonyms for
the same function:
function add(a, b) {
return a+b
}
sum = add
Without the keyword static, the defined function is global and can be accessed
from the whole application. With the keyword static, the function is local to the
current program unit, exactly as if name was declared with the keyword var :
var name = some function value
Default values
Lists the default values used in OPL.
Table 19. Default value syntax
Syntax Effect
with
(expression)
statement
Evaluate expression, then execute statement
with the value of expression temporarily
installed as the default value.
When a reference to an identifier name in
statement is evaluated, this identifier is first
looked up as a property of the default value;
if the default value does not have such a
property, name is treated as a regular
variable.
For example, the following program displays
"The length is 3", because the identifier
length is taken as the length property of the
string "abc".
with ("abc") { writeln("The length is ",
length) }
You can nest with statements; in this case,
references to identifiers are looked up in the
successive default values, from the
innermost to the outermost with statement.
Built-in values and functions
Presents the built-in values and functions of the IBM ILOG Script language for
OPL: numbers, strings, Booleans, arrays, objects, dates, the null and undefined
values, functions.
Numbers
Number representations and functions.
Chapter 3. IBM ILOG Script for OPL 103
Introduction
Provides an overview of how numbers are expressed in OPL.
Numbers can be expressed in decimal (base 10), hexadecimal (base 16), or octal
(base 8.) There are also special numbers.
Note:
For C/C++ programmers: Numbers have the same syntax as C and C++ integers
and doubles. They are internally represented as 64-bit double-precision
floating-point numbers.
Decimal numbers
Explains the use of decimal numbers in OPL.
A decimal number consists of a sequence of digits, followed by an optional
fraction, followed by an optional exponent. The fraction consists of a decimal point
(.) followed by a sequence of digits; the exponent consists of an e or E followed by
an optional + or - sign and a sequence of digits. A decimal number must have at
least one digit.
Here are some examples of decimal number literals:
15
3.14
4e100
.25
5.25e-10
Hexadecimal numbers
Shows how hexadecimal numbers are used in OPL.
A hexadecimal number consists of a 0x or 0X prefix, followed by a sequence of
hexadecimal digits, which include digits 0-9 and the letters a-f or A-F. For example:
0x3ff
0x0
Octal numbers
Explains the use of octal numbers in OPL.
An octal number consists of a 0 followed by a sequence of octal digits, which
include the digits 0-7. For example:
0123
0777
Special numbers
Explains NaN, Infinity, –Infinity, used in OPL.
There are three special numbers: NaN (Not-A-Number), Infinity (positive infinity),
and -Infinity (negative infinity).
NaN
The special number NaN is used to indicate errors in number manipulations. For
example, the square root function Math.sqrt applied to a negative number returns
NaN. There is no representation of NaN as a number literal, but the global script
variable NaN contains its value.
104 OPL Language Reference Manual
The NaN value is contagious, and a numeric operation involving NaN always
returns NaN. A comparison operation involving NaN always returns false — even
the NaN == NaN comparison.
Examples of NaN
Math.sqrt(-1) NaN
Math.sqrt(NaN) NaN
NaN + 3 NaN
NaN == NaN false
NaN <= 3 false
NaN >= 3 false
Note that, according to the ECMA specification http://www.ecma-
international.org/publications/standards/Ecma-262.htm, the operation ToInt32
(NaN) returns the value 0. For example:
int i;
float f;
execute {
i = NaN;
f = NaN;
writeln(i);
writeln(f);
}
returns:
0
NaN
Infinity, –Infinity
The special numbers Infinity and -Infinity are used to indicate infinite values and
overflows in arithmetic operations. The global script variable Infinity contains the
positive infinity. The negative infinity can be computed using the negation
operator (-Infinity).
Examples of Infinity
1/0 Infinity
-1/0 -Infinity
1/Infinity 0
Infinity == Infinity true
Automatic conversion to a number
Describes the automatic conversion of numbers in OPL functions.
When a function or a method which expects a number as one of its arguments is
passed a nonnumeric value, it tries to convert this value to a number using the
following rules:
vA string is parsed as a number literal. If the string does not represent a valid
number literal, the conversion yields NaN.
vThe Boolean true yields the number 1.
vThe Boolean false yields the number 0.
vThe null value yields the number 0.
vA date yields the corresponding number of milliseconds since 00:00:00 UTC,
January 1, 1970.
Chapter 3. IBM ILOG Script for OPL 105
For example, if the Math.sqrt function is passed a string, this string is converted to
the number it represents:
Math.sqrt("25") 5
Similarly, operators which take numeric operands attempt to convert any
nonnumeric operands to a number:
"3" * "4" 12
For operators that can take both strings (concatenation) and numbers (addition),
such as +, the conversion to a string takes precedence over the conversion to a
number (See “Automatic conversion to a string” on page 111.). In other words, if at
least one of the operands is a string, the other operand is converted to a string; if
none of the operands is a string, the operands are both converted to numbers. For
example:
"3" + true "3true"
3 + true 4
For comparison operators, such as == and >=, the conversion to a number takes
precedence over the conversion to a string. In other words, if at least one of the
operands is a number, the other operand is converted to a number. If both
operands are strings, the comparison is made on strings. For example:
"10" > "2" false
"10" > 2 true
Number methods
Provides a reference for the number method in OPL.
There is only one number method.
Table 20. Number method
Syntax Effect
number.toString() Returns a string representing the number as
a literal.
For example:(14.3e2).toString() "1430"
Numeric functions
Provides a reference for numeric functions in the OPL language.
Note:
For C/C++ programmers: Most of the numeric functions are wrap-ups for standard
math library functions.
Table 21. Numeric functions
Syntax Effect
Math.abs(x) Returns the absolute value of x.
Math.max(x,y) Math.max(x, y) returns the larger of x and y.
Math.min(x,y) Math.max(x, y) returns the smaller of x and
y.
Math.random() Returns a pseudo-random number between
0, inclusive, and 1, exclusive.
106 OPL Language Reference Manual
Table 21. Numeric functions (continued)
Syntax Effect
Math.ceil(x) Math.ceil(x) returns the smallest integer
value greater than or equal to x.
Math.floor(x) Math.floor(x) returns the greatest integer
value less than or equal to x.
Math.round(x) Math.round(x) returns the nearest integer
value to x.
Math.sqrt(x) Returns the square root of x.
Math.sin(x) Math.sin(x) returns the trigonometric
function sine of a radian argument.
Math.cos(x) Math.cos(x) returns the trigonometric
function cosine of a radian argument.
Math.tan(x) Math.tan(x) returns the trigonometric
function tangent of a radian argument.
Math.asin(x) Math.asin(x) returns the arcsine of x in the
range -pi/2 to pi/2.
Math.acos(x) Math.acos(x) returns the arc cosine of x in
the range 0to pi.
Math.atan(x) Math.atan(x) returns the arc tangent of x in
the range -pi/2 to pi/2.
Math.atan2(y,x) Math.atan2(y, x) converts rectangular
coordinates (x, y) to polar coordinates (r,
a) by computing aas an arc tangent of y/x
in the range -pi to pi.
Math.exp(x) Math.exp(x) computes the exponential
function.
Math.log(x) Math.log(x) computes the natural logarithm
of x.
Math.pow(x,y) Math.pow(x, y) computes x raised to the
power y.
Numeric constants
Provides a reference for numeric constants in OPL.
The following numeric constants are defined.
Table 22. Numeric constants
Syntax Value
NaN Contains the NaN value.
Infinity Contains the Infinity value.
Number.NaN Same as NaN.
Number.MAX_VALUE The maximum representable number,
approximately 1.79E+308.
Number.MIN_VALUE The smallest representable positive number,
approximately 2.22E-308.
Math.E Napier’s constant, e, and the base of natural
logarithms, approximately 2.718.
Chapter 3. IBM ILOG Script for OPL 107
Table 22. Numeric constants (continued)
Syntax Value
Math.LN10 The natural logarithm of 10, approximately
2.302.
Math.LN2 The natural logarithm of 2, approximately
0.693.
Math.LOG2E The base 2 logarithm of e, approximately
1.442.
Math.LOG10E The base 10 logarithm of e, approximately
0.434.
Math.PI The ratio of the circumference of a circle to
its diameter, approximately 3.142.
Math.SQRT1_2 The square root of one-half, approximately
0.707.
Math.SQRT2 The square root of 2, approximately 1.414.
Numeric operators
Provides a reference for numeric operators in OPL.
Note:
For C/C++ programmers: The numeric operators are the same as in C and C++.
Table 23. Numeric operators
Syntax Effect
x+y
x-y
x*y
x/y
The usual arithmetic operations.
Examples:
3 + 4.2 –> 7.2
100 - 120 –> -20
4 * 7.1 –> 28.4
6/5–>1.2
-x Negation.
Examples:
- 142
-142
x%y Returns the floating-point remainder of
dividing x by y.
Examples:
12%5–>2
12.5%5–>2.5
108 OPL Language Reference Manual
Table 23. Numeric operators (continued)
Syntax Effect
x==y
x!=y
The operator == returns true if x and y are
equal, and false otherwise. The operator !=
is the converse of ==.
Examples:
12 == 12 –> true
12 == 12.1 –> false
12 != 12.1 –> true
x<y
x<=y
x>y
x>=y
The operator <returns true if x is less than
y, and false otherwise. The operator <=
returns true if x is less than or equal to y,
and false otherwise; and so on.
Examples:
-1<0–>true
1<1–>false
1<=1–>true
x&y
x|y
x^y
The bitwise operations & (AND), |
(inclusive OR), and ^ (exclusive OR), where
x and y must be integers in the range
-2**31+1 to 2**31-1 (-2147483647 to
2147483647.)
Examples:
14&9–>8(because 1110 & 1001 –> 1000)
14|9–>15(because 1110 | 1001 –> 1111)
14^9–>7(because 1110 ^ 1001 –> 111)
~x The bitwise NOT operation, where x must
be an integer in the range -2**31+1 to 2**31-1
(-2147483647 to 2147483647.)
Examples:
~ 14 –> 1 (because ~ 1110 –> 0001)
Chapter 3. IBM ILOG Script for OPL 109
Table 23. Numeric operators (continued)
Syntax Effect
x<<y
x>>y
x>>>y
Binary shift operations, where x and y must
be integers in the range -2**31+1 to 2**31-1
(-2147483647 to 2147483647.) The operator <<
shifts to the left, >> shifts to the right
(maintaining the sign bit). The left operand
specifies the value to be shifted. The right
operand specifies the number of positions
that the bits in the value are to be shifted.
>>> shifts to the right, shifting in zeros from
the left.
Examples:
9 << 2 –> 36 (because 1001 << 2 –> 100100)
9 >> 2 –> 2 (because 1001 >> 2 –> 10)
-9 >> 2 –> -2 (because 1..11001 >> 2 –>
1..11110)
-9 >>> 2 –> 1073741821 (because 1..11001
>>> 2 –> 01..11110)
IBM ILOG Script strings
String representation and functions.
Introduction
Provides an overview of the use of strings in IBM ILOG Script.
A string literal is zero or more characters enclosed in double (“) or single (’)
quotes.
Note:
For C/C++ programmers: Except for the use of single quotes, string literals have
the same syntax as in C and C++.
Here are examples of string literals:
"My name is Hal"
'My name is Hal'
'"Hi there", he said'
"3.14"
"Hello, world\n"
In these examples, the first and second strings are identical.
The backslash character \can be used to introduce an escape sequence, which
stands for a character which cannot be directly expressed in a string literal.
Table 24. Escape sequences in strings
Escape sequence Stands for
\n Newline
\t Tab
110 OPL Language Reference Manual
Table 24. Escape sequences in strings (continued)
Escape sequence Stands for
\\ Backslash character (\)
\” Double quote (")
\’ Single quote (’)
\b Backspace
\f Form feed
\r Carriage return
\xhh The character whose ASCII code is hh,
where hh is a sequence of two hexadecimal
digits.
\ooo The character whose ASCII code is ooo,
where ooo is a sequence of one, two, or
three octal digits.
Here are examples of string literals using escape sequences:
Table 25. Examples of string literals using escape sequences
String literal Stands for
"Read ’The Black Bean’" Read ‘The Black Bean’
'"Hello", he said' “Hello”, he said
"c:\\temp" c:\temp
"First line\nSecond line\nThird line" First line
Second line
Third line
"\xA9 1995-1997" © 1995-1997
When a string is converted to a number, an attempt is made to parse it as a
number literal. If the string does not represent a valid number literal, the
conversion yields NaN.
Automatic conversion to a string
Explains how nonstrings are automatically converted to strings.
When a function or a method that expects a string as one of its arguments is
passed a nonstring value, this value is automatically converted to a string. For
example, if the string method indexOf is passed a number as its first argument,
this number is treated like its string representation:
"The 10 commandments".indexOf(10) -> 4
Similarly, operators that take string operands automatically convert nonstring
operands to strings:
"The"+10+"commandments" -> "The 10 commandments"
The conversion to a string uses the toString method of the given value. All
built-in values have a toString method.
Chapter 3. IBM ILOG Script for OPL 111
String properties
Provides a reference for the properties of strings.
There is a single, read-only string property.
Table 26. String property
Syntax Value
string.length Number of characters in string. This is a
read-only property.
Examples:
"abc".length –> 3
"".length –> 0
String methods
Provides a reference for string methods.
Characters in a string are indexed from left to right. The index of the first character
in a string is 0, and the index of the last character is string.length-1.
Table 27. String methods
Syntax Effect
string.substring(start [,end ] ) Returns the substring of string starting at
the index start and ending at the index
end-1.Ifend is omitted, the tail of string is
returned.
Examples:
"0123456".substring(0, 3) –> "012"
"0123456".substring(2, 4) –> "23"
"0123456".substring(2) –> "23456"
string.charAt(index) Returns a one-character string containing the
character at the specified index of string. If
index is out of range, an empty string is
returned.
Examples:
"abcdef".charAt(0) –> "a"
"abcdef".charAt(3) –> "d"
"abcdef".charAt(100) –> ""
string.charCodeAt(index) Returns the ASCII code of the character at
the specified index of string.Ifindex is out
of range, returns NaN.
Examples:
"abcdef".charCodeAt(0) –> 97
"abcdef".charCodeAt(3) –> 100
"abcdef".charCodeAt(100) –> NaN
112 OPL Language Reference Manual
Table 27. String methods (continued)
Syntax Effect
string.indexOf(substring [,index ] ) Returns the index in string of the first
occurrence of substring. The string is
searched starting at index.Ifindex is
omitted, string is searched from the
beginning. This method returns -1 if
substring is not found.
Examples:
"abcdabcd".indexOf("bc") –> 1
"abcdabcd".indexOf("bc", 1) –> 1
"abcdabcd".indexOf("bc", 2) –> 5
"abcdabcd".indexOf("bc", 10) –> -1
"abcdabcd".indexOf("foo") –> -1
"abcdabcd".indexOf("BC") –> -1
string.lastIndexOf(substring [,index ] ) Returns the index in string of the last
occurrence of substring, when string is
searched backwards, starting at index.If
index is omitted, string is searched from the
end. This method returns -1 if substring is
not found.
Examples:
"abcdabcd".lastIndexOf("bc") –> 5
"abcdabcd".lastIndexOf("bc", 5) –> 5
"abcdabcd".lastIndexOf("bc", 4) –> 1
"abcdabcd".lastIndexOf("bc", 0) –> -1
"abcdabcd".lastIndexOf("foo") –> -1
"abcdabcd".lastIndexOf("BC") –> -1
string.toLowerCase() Returns string converted to lowercase.
Example: "Hello, World".toLowerCase()
"hello, world"
string.toUpperCase() string.toUpperCase() Returns string
converted to uppercase.
Example:"Hello, World".toUpperCase()
"HELLO, WORLD"
Chapter 3. IBM ILOG Script for OPL 113
Table 27. String methods (continued)
Syntax Effect
string.split(separator) Returns an array of strings containing the
substrings of string that are separated by
separator. See also the array method join.
Examples:
"first name,last name,age".split(",") ->
an array asuch that .length is 3, a[0] is
"first name",a[1] is "last name", and a[2]
is "age".
If string does not contain separator,an
array with one element containing the whole
string is returned.
Example:
"hello".split(",") –> an array a such that
a.length is 1 and a[0] is "hello",
string.toString() Returns the string itself.
String functions
Provides a reference of string functions.
Table 28. String functions
Syntax Effect
String.fromCharCode(code) Returns a single-character string containing
the character with the given ASCII code.
Examples:
String.fromCharCode(65) –>"A"
writeln(String.fromCharCode(0x30)); –> "0"
114 OPL Language Reference Manual
Table 28. String functions (continued)
Syntax Effect
parseInt(string [,base ] ) Parses string as an integer written in the
given base, and returns its value. If the
string does not represent a valid integer,
NaN is returned.
Leading white space characters are ignored.
If parseInt encounters a character that is not
a digit in the specified base, it ignores it and
all succeeding characters and returns the
integer value parsed up to that point.
If base is omitted, it is taken to be 10, unless
string starts with 0x or 0X, in which case it
is parsed in base 16, or with 0, in which case
it is parsed in base 8.
Examples:
parseInt("123") –> 123
parseInt("-123") –> -123
parseInt("123.45") –> 123
parseInt("1001010010110", 2) –> 4758
parseInt("a9", 16) –> 169
parseInt("0xa9") –> 169
parseInt("010") –> 8
parseInt("123 poodles") –> 123
parseInt("a lot of poodles") –> NaN
parseFloat(string) Parses string as a floating-point number and
return its value. If the string does not
represent a valid number, NaN is returned.
Leading white space characters are ignored.
The string is parsed up to the first
unrecognized character. If no number is
recognized, the function returns NaN.
Examples:
parseFloat("-3.14e-15") –> -3.14e-15
parseFloat("-3.14e-15 poodles") –> -3.14e-15
parseFloat("a fraction of a poodle") –> NaN
Chapter 3. IBM ILOG Script for OPL 115
String operators
Provides a reference of string operators.
Table 29. String operators
Syntax Effect
string1 +string2 Returns a string containing the
concatenation of string1 and string2.
Examples:"Hello," + " world" –> "Hello,
world"
When the operator + is used to add a string
to a nonstring value, the nonstring value is
first converted to a string.
Examples:
"Yourageis"+23–>"Yourageis23"
23+"isyour age" –> "23 is your age"
string1 == string2
string1 != string2
The operator == returns the Boolean true if
string1 and string2 are identical, and false
otherwise. Two strings are identical if they
have the same length and contain the same
sequence of characters. The operator != is
the converse of ==.
Examples:
"a string" == "a string" –> true
"a string" == "another string" –> false
"a string" == "A STRING" –> false
"a string" != "a string" –> false
"a string" != "another string" –> true
When the operators == and != are used to
compare a string with a number, the string
is first converted to a number and the two
numbers are compared numerically.
Examples:
"12" == "+12" –> false
12 == "+12" –> true
116 OPL Language Reference Manual
Table 29. String operators (continued)
Syntax Effect
string1 <string2
string1 <= string2
string1 >string2
string1 >= string2
The operator <returns true if string1
strictly precedes string2 lexicographically,
and false otherwise. The operator <= returns
true if string1 strictly precedes string2
lexicographically or is equal to it, and false
otherwise; and so on.
Examples:
"abc" < "xyz" –> true
"a" < "abc" –> true
"xyz" < "abc" –> false
"abc" < "abc" –> false
"abc" > "xyz" –> false
"a" > "abc" –> false
"xyz" > "abc" –> true
Etc.
When one of these operators is used to
compare a string with a number, the string
is first converted to a number and the two
numbers are compared numerically. In all
other cases, the other argument is first
converted to a string.
Examples:
"10" > "2" –> false
10 > "2" –> true
123 < "2" –> false
Hint: Autocasting may cause unexpected
behavior.
IBM ILOG Script Booleans
Boolean representation and functions.
Introduction
Provides an overview of the use of Boolean literals in IBM ILOG Script.
There are two Boolean literals: true, which represents the Boolean value true, and
false, which represents the Boolean value false.
When converted to a number, true yields 1 and false yields 0.
Automatic conversion to Boolean
Describes the automatic conversion of non-Booleans to Booleans.
Chapter 3. IBM ILOG Script for OPL 117
When a function, method or statement which expects a Boolean value as one of its
arguments is passed a non-Boolean value, this value is automatically converted to
a Boolean value as follows:
vThe number 0yields false.
vThe empty string "" yields false.
vThe null value yields false.
vThe undefined value yields false.
vAny other non-Boolean values yield true.
For example:
if ("") writeln("True"); else writeln("False");
if (123) writeln("True"); else writeln("False");
This displays “False”, then “True”.
Boolean methods
Provides a reference of Boolean methods.
There is only one Boolean method.
Table 30. Boolean method
Syntax Effect
boolean.toString() Returns a string representing the Boolean
value, either "true" or "false".
Example:
true.toString –> "true"
false.toString –> "false"
Logical operators
Provides a reference of logical operators.
The following Boolean operators are available.
Note:
For C/C++ programmers: These operators are the same as in C and C++.
Table 31. Logical operators
Syntax Effect
!boolean Logical negation.
Examples:
! true –> false
! false –> true
118 OPL Language Reference Manual
Table 31. Logical operators (continued)
Syntax Effect
exp1 && exp2 Logical AND. Returns true if both Boolean
expressions exp1 and exp2 are true.
Otherwise, returns false.
If exp1 is false, this expression immediately
returns false without evaluating exp2,so
any side effects of exp2 are not taken into
account.
Examples:
true && true –> true
true && false –> false
false && whatever –> false (whatever is not
evaluated)
exp1 || exp2 Logical OR. Returns true if either Boolean
expression exp1 or exp2 (or both) is true.
Otherwise, returns false.
If exp1 is true, this expression immediately
returns true without evaluating exp2, so
any side effects of exp2 are not taken into
account.
Examples:
false || true –> true
false || false –> false
true || whatever –> true (whatever is not
evaluated)
condition ?exp1 :exp2 If condition is true, this expression returns
exp1 ; otherwise, it returns exp2.
When condition is true, the expression exp2
is not evaluated, so any side effects it may
contain are not taken into account. Similarly,
when condition is false,exp1 is not
evaluated.
Examples:
true ? 3.14 : whatever –> 3.14
false ? whatever : "Hello" –> "Hello"
IBM ILOG Script arrays
Array representation and functions.
Introduction
Provides an overview of the use of arrays in IBM ILOG Script.
Arrays provide a way of manipulating ordered sets of values referenced through
an index starting from zero (0). Unlike arrays in other languages, IBM ILOG Script
Chapter 3. IBM ILOG Script for OPL 119
arrays do not have a fixed size and are automatically expanded as new elements
are added. For example, in the following program, an array is created empty, and
new elements are then added.
a = new Array() // Create an empty array
a[0] = "first" // Set the element 0
a[1] = "second" // Set the element 1
a[2] = "third" // Set the element 2
Arrays are internally represented as sparse objects, which means that an array
where only the element 0 and the element 10000 have been set occupies just
enough memory to store these two elements, not the 9999 which are between 0
and 10000.
Array constructor
Provides a reference for array constructors in IBM ILOG Script.
The array constructor has two distinct forms.
Table 32. Array constructor
Syntax Effect
new Array(length) Returns a new array a of length length with
its elements from 0to length-1 set to null.
If length is not a number, and its conversion
to a number yields NaN, the second syntax
is used.
Examples:
new Array(12) –> an array a with length 12
and a[0] to a[11] containing null
new Array("5") –> an array a with length 5
and a[0] to a[4] containing null
new Array("foo") see second syntax
new Array(element1, ..., elementn) Returns a new array a of length n with a[0]
containing element1,a[1] containing
element2, and so on. If no argument is
given, that is n=0, an empty array is created.
If n=1 and element1 is a number or can be
converted to a number, the first syntax is
used.
Examples:
new Array(327, "hello world") –> an array a
of length 2 with a[0] == 327 and a[1] ==
"hello world"
new Array() –> an array with length 0
new Array("327") see first syntax
120 OPL Language Reference Manual
Array properties
Provides a reference of the properties of arrays.
Table 33. Array properties
Syntax Effect
array[index]Ifindex can be converted to a number
between 0 and 2e31-1 (see “Automatic
conversion to a number” on page 105),
array[index] is the value of the index-th
element of the array.
Otherwise, it is considered as a standard
property access.
If this element has never been set, null is
returned.
Example:
Suppose that the array ahas been created
witha = new Array("foo", 12, true)
Then:
a[0] –> "foo"
a[1] –> 12
a[2] –> true
a[3] –> null
a[1000] –> null
When an element of an array is set beyond
the current length of the array, the array is
automatically expanded:a[1000] = "bar" //
the array is automatically expanded.
Unlike other properties, the numeric
properties of an array are not listed by the
for..in statement.
Chapter 3. IBM ILOG Script for OPL 121
Table 33. Array properties (continued)
Syntax Effect
array.length The length of array, which is the highest
index of an element set in array, plus one. It
is always included in 0 to 2e31 -1.
When a new element is set in the array, and
its index is greater than or equal to the
current array length, the length property is
automatically increased.
Example: Suppose that the array ahas been
created witha = new Array("a", "b", "c")
Then:a.length -> 3 a[100] = "bar"; a.length –>
101
You can also change the length of an array
by setting its length property.
a = new Array(); a[4] = "foo"; a[9] = "bar"
a.length –> 10
a.length = 5 a.length –> 5 a[4] –> "foo" a[9]
–> null
Array methods
Provides a reference of array methods.
Table 34. Array methods
Syntax Effect
array.join([ separator ]) Returns a string that contains the elements
of the array converted to strings,
concatenated together and separated with
separator.Ifseparator is omitted, it is taken to
be ",". Elements that are not initialized are
converted to the empty string. See also the
string method split.
Example: Suppose that the array ahas been
created with
a = new Array("foo", 12, true)
Then:
a.join("//") –> "foo//12//true"
a.join() –> "foo,12,true"
122 OPL Language Reference Manual
Table 34. Array methods (continued)
Syntax Effect
array.sort([ function ]) Sorts the array. The elements are sorted in
place; no new array is created.
If function is not provided, array is sorted
lexicographically: Elements are compared by
converting them to strings and using the <
operator. With this order, the number 20
would come before the number 5, since "20"
< "5" is true.
If function is supplied, the array is sorted
according to the return value of this
function. This function must take two
arguments x and y and return:
-1 if x is smaller than y
0if x is equal to y
1if x is greater than y
Example: Suppose that the function
compareLength is defined as
function compareLength(x, y) {if (x.length <
y.length) return -1; else if (x.length ==
y.length) return 0; else return 1; }
and that the array ahas been created with:
a = new Array("giraffe", "rat",
"brontosaurus")
Then a.sort() will reorder its elements as
follows:"brontosaurus" "rat" "giraffe"
while a.sort(compareLength) will reorder
them as follows:"rat" "giraffe" "brontosaurus"
array.reverse() Transposes the elements of the array: the
first element becomes the last, the second
becomes the second to last, and so on. The
elements are reversed in place; no new array
is created.
Example: Suppose that the array ahas been
created with
a = new Array("foo", 12, "hello", true, false)
Then a.reverse() changes aso that:
a[0] false
a[1] true
a[2] "hello"
a[3] 12
a[4] "foo"
Chapter 3. IBM ILOG Script for OPL 123
Table 34. Array methods (continued)
Syntax Effect
array.toString() Returns the string "[object Object]".
Objects
Object representation and functions.
Introduction
Provides an overview of the use of objects in IBM ILOG Script.
Objects are values which do not contain any predefined properties or methods
(except the toString method), but where new ones can be added. A new, empty
object can be created using the Object constructor. For example, the following
program creates a new object, stores it in the variable myCar, and adds the
properties “name” and “year” to it:
myCar = new Object() // o contains no properties
myCar.name = "Ford"
myCar.year = 1985
Now:
myCar.name -> "Ford"
myCar.year -> 1985
Defining methods
Explains how methods are defined in IBM ILOG Script.
Since a method is really a property which contains a function value, defining a
method simply consists in defining a regular function, then assigning it to a
property.
For example, the following program adds a method start to the myCar object
defined in the previous section:
function start_engine() {
writeln("vroom vroom")
}
myCar.start = start_engine
Now, the expression myCar.start() will call the function defined as start_engine.
Note that the only reason for using a different name for the function and for the
method is to avoid confusion; we could have written:
function start() {
writeln("vroom vroom")
}
myCar.start = start
this as a keyword
Describes the use of the this keyword in IBM ILOG Script.
Inside methods, the this keyword can be used to reference the calling object. For
example, the following program defines a method getName, which returns the value
of the name property of the calling object, and adds this method to myCar :
124 OPL Language Reference Manual
function get_name() {
return this.name
}
myCar.getName = get_name
Inside constructors, this references the object created by the constructor.
When used in a nonmethod context, this returns a reference to the global object.
The global object contains script variables declared at top level, and built in
functions and constructors.
Object constructor
Provides a reference of object constructors in IBM ILOG Script.
Table 35. Object constructor
Syntax Effect
new Object() Returns a new object with no properties.
User-defined constructors
Provides a reference of user-defined constructors.
In addition to the Object constructor, any user-defined function can be used as an
object constructor.
Table 36. User-defined constructors
Syntax Effect
new function(arg1, ..., argn) Creates a new object, then
calls function(arg1, ..., argn) to initialize
it.
Inside the constructor, the keyword this can be used to make reference to the
object being initialized.
For example, the following program defines a constructor for cars:
function Car(name, year) {
this.name = name;
this.year = year;
this.start = start_engine;
}
Now, calling
new Car("Ford", "1985")
creates a new object with the properties name and year, and a start method.
Built-in methods
Provides a reference of the built-in methods of IBM ILOG Script.
There is only one object built-in method.
Table 37. Built-in method
Syntax Effect
object.toString() Returns the string "[object Object]". This
method can be overridden by assigning the
toString property of an object.
Chapter 3. IBM ILOG Script for OPL 125
Dates
Date representation and functions
Introduction
Provides an overview of dates and date functions in IBM ILOG Script.
Date values provide a way of manipulating dates and times. Dates can be best
understood as internally represented by a number of milliseconds since 00:00:00
UTC, January 1, 1970. This number can be negative, to express a date before 1970.
Note:
1. For C/C++ programmers: Unlike dates manipulated by the standard C library,
date values are not limited to the range of 1970 to 2038, but span
approximately 285,616 years before and after 1970.
2. IBM ILOG Script measures time in wall-clock time (and not system time).
When converted to a number, a date yields the number of milliseconds since
00:00:00 UTC, January 1, 1970.
Date constructor
Explains the different forms of the date constructor.
The date constructor has four distinct forms.
Table 38. Date constructor
Syntax Effect
new Date() Returns the date representing the current
time.
new Date(milliseconds) Returns the date representing 00:00:00 UTC,
January 1, 1970, plus milliseconds
milliseconds. The argument can be negative,
to express a date before 1970. If the
argument cannot be converted to a number,
the third constructor syntax is used.
Examples:
new Date(0) -> a date representing 00:00:00
UTC, January 1, 1970.
new Date(1000*60*60*24*20) -> a date
representing twenty days after 00:00:00 UTC,
January 1, 1970.
new Date(-1000*60*60*24*20) -> a date
representing twenty days before 00:00:00
UTC, January 1, 1970.
126 OPL Language Reference Manual
Table 38. Date constructor (continued)
Syntax Effect
new Date(string) Returns the date described by string,
which must have the form:
month/day/year hour:minute:second
msecond
The date expressed in string is taken in
local time.
Example:
new Date("12/25/1932 14:35:12 820")
A date representing December 25th, 1932, at
2:35 PM plus 12 seconds and 820
milliseconds, local time.
new Date(year,
month,
[,day
[,hours
[,minutes
[,seconds
[,mseconds ]]]]])
Returns a new date representing the given
year, month, day, and so on, taken in local
time. The arguments are:
year: any integer.
month: range 0-11 (where 0=January,
1=February, and so on)
day: range 1-31, default 1
hours: range 0-23, default 0
minutes: range 0-59, default 0
seconds: range 0-59, default 0
mseconds: range 0-999, defaults to 0
Examples:
new Date(1932, 11, 25, 14, 35, 12, 820)
A date representing December 25th, 1932, at
2:35 PM plus 12 seconds and 820
milliseconds, local time.
new Date(1932, 11, 25)
A date representing December 25th, 1932, at
00:00, local time.
Chapter 3. IBM ILOG Script for OPL 127
Date methods
Provides a reference of date methods in IBM ILOG Script.
Table 39. Date methods
Syntax Effect
date.getTime()
date.setTime(milliseconds)
Returns (or sets) the number of milliseconds
since 00:00:00 UTC, January 1, 1970.
Example: Suppose that the date dhas been
created with:
d = new Date(3427)
Then:
d.getTime() –> 3427
date.toLocaleString()
date.toUTCString()
Returns a string representing the date in
local time or in UTC respectively.
Example: Suppose that the date dhas been
created with:
d = new Date("3/12/1997 12:45:00 0")
Then:
d.toLocaleString() –> "03/12/1997 12:45:00
000"
d.toUTCString() –> "03/12/1997 10:45:00
000"
assuming a local time zone offset of +2
hours with respect to Greenwich Mean Time.
date.getYear()
date.setYear(year)
Returns (or sets) the year of date.
date.getMonth()
date.setMonth(month)
Returns (or sets) the month of date.
date.getDate()
date.setDate(day)
Returns (or sets) the day of date.
date.getHours()
date.setHours(day)
Returns (or sets) the hours of date.
date.getMinutes()
date.setMinutes(minutes)
Returns (or sets) the minutes of date.
date.getSeconds()
date.setSeconds(seconds)
Returns (or sets) the seconds of date.
date.getMilliseconds()
date.setMilliseconds(millisecs)
Returns (or sets) the milliseconds of date.
date.toString() Returns the same value as
date.toLocaleString()
Related reference:
128 OPL Language Reference Manual
“Tips on date formats”
Provides some tips on using ILOG Script to improve the date format.
Date functions
Provides a reference of date functions.
Table 40. Date functions
Syntax Effect
Date.UTC(year, month, [ , day [ , hours [ ,
minutes [ , seconds [ , mseconds ]]]]])
Returns a number representing the given
date taken in UTC. The arguments are:
year: any integer
month: range 0-11, where 0 = January, 1 =
February, and so on
day: range 1-31, default 1
hours: range 0-59, default 0
minutes: range 0-59, default 0
seconds: range 0-59, default 0
mseconds: range 0-999, default 0
Date.parse(string) Same as new Date(string), but the result is
returned as a number rather than as a date
object.
Date operators
Explains the use of date operators in IBM ILOG Script.
There are no specific operators for dealing with dates, but, since numeric operators
automatically convert their arguments to numbers, these operators can be used to
compute the time elapsed between two dates, to compare dates, or to add a given
amount of time to a date. For example:
date1 - date2 -> the number of milliseconds elapsed between date1 and date2.
date1 < date2 -> true if date1 is before date2, false otherwise.
new Date(date+10000) ->
a date representing 10000 milliseconds after date.
The following program displays the number of milliseconds taken to execute the
statement <do something> :
before = new Date();
<do something>;
after = new Date();
writeln("Time for doing something: ", after-before, " milliseconds.");
Tips on date formats
Provides some tips on using ILOG Script to improve the date format.
Setting the date
The set methods on dates perform safety checks. For example, when you try to set
the day, the corresponding month must have the correct number of days. If the
date is 2011/ February/15 and you try to change the day to 31, an error will occur.
Therefore, to avoid errors when you create a new date, you must respect the order:
Year / Month / Day.
Chapter 3. IBM ILOG Script for OPL 129
Modifying the date format
If your date is not in the format you want, you need to use ILOG Script to modify
it. Below is an example.
execute {
function converter(dateIn) {
dateTime = dateIn.split(" ");
date = dateTime[0]; time = dateTime[1];
dateSplit = date.split("-");
dateReverse = dateSplit.reverse();
dateNew = dateReverse.join("/")+""+time;
return Date.parse(dateNew);
}
var d1 = converter("2011-07-06 07:00:00.0");
writeln(d1);
}
This code will return a new date in the format day/month/year. So in the
example, the date returned is:
06/07/2011 07:00:00.0
The null value
Explains the use of the null value.
The null value is a special value used in some places to specify an absence of
information.
For example, an array element which has not yet been set has a default value of
null. The null value is not to be confused with the undefined value, which also
specifies an absence of information in some contexts. See section “The undefined
value” below.
The null value can be referenced in programs with the keyword null :
null -> the null value
When converted to a number, null yields zero (0).
Methods of null
There is only one method of null.
Table 41. Methods of null
Syntax Effect
null.toString() Returns the string "null".
The undefined value
Explains the use of the undefined value.
The undefined value is a special value used in some places to specify an absence
of information. For example, accessing a property of a value which is not defined,
or a local variable which has been declared but not initialized, yields the undefined
value.
130 OPL Language Reference Manual
There is no way of referencing the undefined value in programs. Checking if a
value is the undefined value can be done using the typeof operator:
typeof(value) == "undefined" -> true if value is undefined,
false otherwise.
Methods of undefined
There is only one method of undefined.
Table 42. Methods of undefined
Syntax Effect
undefined.toString() Returns the string "undefined".
IBM ILOG Script functions
Describes the use of functions in IBM ILOG Script.
In IBM ILOG Script, functions are regular values (also known as “first class”
values) which can be manipulated like any other type of value: They can be passed
to functions, returned by functions, stored into script variables or into object
properties, and so on.
For example, the function parseInt is a function value which is stored in the
parseInt variable:
parseInt ->
a function value
This function value can be, for example, assigned to another variable:
myFunction = parseInt
and then called through this variable:
myFunction("-25") -> -25
Function methods
There is only one method of functions.
Table 43. Function methods
Syntax Effect
function.toString() Returns a string which contains some
information about the function.
Examples:
"foo".substring.toString() "[primitive method
substring]"
eval.toString() "[primitive function eval]"
Chapter 3. IBM ILOG Script for OPL 131
Miscellaneous functions
Provides a reference of miscellaneous functions in IBM ILOG Script.
Table 44. Miscellaneous functions
Syntax Effect
stop() Stops the execution of the program at the
current statement and, if the debugger is
enabled, enters debug mode.
write (arg1, ..., argn)
writeln (arg1, ..., argn)
Converts the arguments to strings and prints
them to the current debug output. The
implementation depends on the application
in which IBM ILOG Script is embedded. The
function writeln prints a newline at the end
of the output, while write does not.
loadFile(string) Loads the script file whose path is string.
The path can be either absolute or relative. If
this path does not designate an existing file,
the file is looked up using a method that
depends on the application in which the
script is embedded. Typically, a file with the
name string is searched for in a list of
directories specified in the application setup.
includeScript(string)
For use in OPL.
Loads the script file whose path is string.
The path can be either absolute or relative. If
this path does not designate an existing file,
the file is looked up using a method that
depends on the application in which the
script is embedded. Typically, a file with the
name string is searched for in a list of
directories specified in the application setup.
The effect of this function is the same as for
loadFile. In OPL, you have to use
includeScript, as the function loadFile
does not recognize OPL objects, such as
thisOplModel.
eval(string) Executes string as a program, and returns
the value of the last evaluated expression.
The program in string can use all the
features of the language, except that it
cannot define functions; in other words, the
function statement is not allowed in string.
Examples:
eval("2*3") -> 6
eval("var i=0; for (var j=0; j<100; j++) i=i+j;
i") -> 4950 n=25; eval("Math.sqrt(n)") -> 5
eval("function foo(x) { return x+1 }") -> error
132 OPL Language Reference Manual
Table 44. Miscellaneous functions (continued)
Syntax Effect
fail() Stops the execution of the scripting block at
the current statement, reports an error, and
goes on.
Example:
execute b1 { writeln("A"); fail(); writeln("B");
} execute b2 {writeln("C"); }
gives
AC
as the output
and reports an error line 4
Scripting runtime error: fail() called.
Chapter 3. IBM ILOG Script for OPL 133
134 OPL Language Reference Manual
Index
A
abs, OPL function 42
accessing
value of an IBM ILOG Script
property 93
accessing named ranges
in Excel 30
aggregate operators 43, 66
all, OPL keyword 39
and, logical constraint 60
arguments
IBM ILOG Script keyword 96
arrays
appending 40
constructor (IBM ILOG Script) 120
explicit 40
in IBM ILOG Script 119
initialization 17
methods (IBM ILOG Script) 122
multidimensional 8
of decision variables, initialization 40
one-dimensional 8
properties (IBM ILOG Script) 121
assert, OPL keyword 34
processing order of statements 35
assertions 34
assignment operators 94
B
Boolean expressions 50
constraints 51
Boolean literals 117
conversion to 118
Boolean method toString 118
brackets, delimiters in IBM ILOG Script
for OPL 91
break, IBM ILOG Script keyword 99
breakpoint 45
breakpoints
in piecewise-linear functions 43
building blocks 3
C
card, OPL function 49
cardinality constraints 61
case-sensitivity of the scripting
language 92
collections
not sorted in tuple sets without
keys 14
comments
delimiters 92
compatibility constraints in CP 63
compound statements
in IBM ILOG Script for OPL 91
conditional constraints 52
conditional expressions
for float and integers 42
conditional statements
in IBM ILOG Script for OPL 91, 99
connection
to a database 24
to a spreadsheet 29
consistency of model data 32
constants 107
dynamic collection 39
constraint labeling
limitations 56
constraints
basic 58
conditional 52
declaration 51
discrete 58
element expressions (CP) 62
float 58
for compatibility (CP) 63
for filtering 52
labeling 53
limitations 64
logical (CPLEX) 59
logical, for CP 62
nonlinear, rejected by CPLEX 58
scheduling 63
specialized (CP) 63
string 59
types 58
using 51
constraints, limitations 64
constructors
for IBM ILOG Script arrays 120
for IBM ILOG Script dates 126
for IBM ILOG Script objects 125
user-defined in IBM ILOG Script 125
continue, IBM ILOG Script keyword 99
conventions
in models 3
conversion (IBM ILOG Script)
of non-Boolean value to a
Boolean 118
of nonnumeric value to a
number 105
of nonstring value to a string 111
costs
discontinuous (pwl) 46
count, OPL function 42
D
data
assertions for consistency 34
consistency 32
initializing 15
input/output
to/from a database 24
to/from a spreadsheet 29
preprocessing 35
reading from an Excel
spreadsheet 29
data files
syntax for databases 24
data initialization, lazy 16, 23, 36
data structures
arrays 8
ranges 7
sets 12
data types
arrays 8
floats 4
integers 4
piecewise linear functions 5
ranges 7
sets 12
stepwise functions 6
strings 5
data types allowed in tuples 11
data, pushing and pulling 23
database
connection strings 25
databases
connecting to 24
data input/output 24
deleting elements 28
reading from 26
SQL encryption 27
writing to 27
date format, tips using script 129
dates in IBM ILOG Script 126
constructor 126
functions 129
methods 128
operators 129
DBExecute, OPL keyword 27
DBRead, OPL keyword 26
DBUpdate, OPL keyword 28
decimal numbers 104
decision expressions 39
and tuple patterns/tuple indices 11
decision variables 37
and integer ranges 7
arrays of, initialization 40
definition 37
dynamic collection 39
reusable (dexpr) 39
declaration
of constraints 51
of script variables
inside a function definition 101
outside a function definition 101
of tuples, using keys 10
default values
of statements in IBM ILOG
Script 103
delimiters
and internal initialization
tuples 21
curly brackets 91
for comments 92
for tuples 20
quotes 110, 117
© Copyright IBM Corp. 1987, 2015 135
delimiters (continued)
semi-colon 91
dexpr, OPL keyword 39
diff, OPL keyword 49
discontinuous piecewise-linear
functions 45
discrete
constraints 58
data 58
variables 58
dvar, OPL keyword
vs. var 37
E
efficient models 53, 66, 67
element expressions in CP 62
else
See if-then-else 52
else, IBM ILOG Script keyword 99
encryption of SQL statements 27
equivalence, logical constraint 60
errors
tuples with same names 10
eval function 132
Excel
accessing named ranges 30
Excel spreadsheet
what data can be read 29
Excel, R1C1 style not supported 29, 31
execute, IBM ILOG Script block
for preprocessing 35
expressions 92
Boolean 50
counting 42
float 42
in IBM ILOG Script for OPL 91
in logical constraints 59
integer 42
set 49
syntax 93
external data 16
and memory allocation 24
sets 21
F
fail function 132
filtering
formal parameters 64
in tuples of parameters 67
with constraints 52
first, OPL function 49
float constraints 58
float, OPL keyword 4
floats 4
expression 42
functions in OPL 42
for, IBM ILOG Script keyword
loop syntax 99
forall, statements
and constraint labels 56
formal parameters
basic 64
filter expressions 64
tuples 66
function calls 95
function definitions 102
function, IBM ILOG Script keyword 102
functions 131
card 49
first 49
float 42
for dates 129
for strings 114
item 49
last 49
next 49
nextc 49
ord 49
over set expressions 49
prev 49
prevc 49
sign 46
functions in ILOG Script
miscellaneous 132
functions, numeric 106
G
generic sets 21, 66
ground
breakpoints and slopes in
piecewise-linear functions 43
conditions in if-then-else
statements 52
expressions and relations 41
H
hexadecimal numbers 104
I
IBM ILOG Script
compound statements 91
default values of statements 103
syntax 91
identifiers 92
conventions 3
for data and variables 41
if-then-else
conditional constraints 52
if, IBM ILOG Script keyword 99
implication
logical constraint (CP) 62
logical constraint (CPLEX) 60
implicit slicing 67
includeScript function 132
indexed labels 54
infinity 104
infinity, OPL keyword 4, 42
initializing
arrays 17
arrays of decision variables 40
data 15, 16
set of tuples 16
sets 21
tuples 20
input/output
data to/from a database 24
data to/from a spreadsheet 29
integer constant maxint 4
integer expressions 42
integer ranges 7
integers 4
inter, OPL keyword 49
internal data 16
and memory allocation 23
sets 21
item function 49
K
keys in tuple declarations 10
keywords
arguments 96
assert 34
break 99
continue 99
DBExecute 27
DBRead 26
DBUpdate 28
diff 49
dvar 37
else 99
float 4
for 99
function 102
if 99
infinity 4, 42
inter 49
max 43
maxint 42
min 43
null 130
ordered 64
piecewise 43
prod 43
return 102
setof 12
SheetConnection 29
SheetRead 29
special 96
static 102
string 5
sum 43
symdiff 49
this 96, 124, 125
tuple 9
union 49
var 101
with 32
L
labeled assertions 55
labeled constraints 53
last, OPL function 49
lazy initialization of data 16, 23, 35
limitations on constraint labeling 56
limitations on constraints 64
limitations on tuples 11
literals 92
loadFile function 132
logical constraints
definition and extraction 59
for CP 62
136 OPL Language Reference Manual
logical expressions 59
logical operators 118
loops
in IBM ILOG Script 99
M
max, OPL keyword 43
maxint, OPL keyword 4, 42
memory allocation and management
and data initialization 23
memory consumption
and multidimensional arrays 8
methods
built-in, for objects 125
defining for objects 124
for arrays 122
for Booleans 118
for dates 128
for functions 131
for numbers 106
for strings 112
for the null value 130
for the undefined value 131
methods for sets 13
min, OPL keyword 43
models
building 3
connecting to databases 24
conventions 3
efficiency 53
readability 66, 67
multidimensional arrays 8
N
NaN (Not-a-Number) 104
next, OPL function 49
nextc, OPL function 49
nonlinear constraints 58
nonlinear expressions in logical
constraints 60
Not-a-Number (NaN) 104
not, logical constraint 60
null value 130
null, IBM ILOG Script keyword 130
numbers 104
automatic conversion to 105
methods 106
numeric constants 107
numeric functions 106
numeric operators 108
O
objective function
and decision variables 37
objects 124
constructor 125
octal numbers 104
one-dimensional arrays 8
operations on sets 12
operators
aggregate 43
assigning a value 94
for IBM ILOG Script dates 129
operators (continued)
for strings 116
logical 118
numeric 108
precedence 93
shorthand 94
special 96
syntax 97
or, logical constraint 60
ord, OPL function 49
order
for processing script blocks 35
ordered sets 12
special ordered sets, not
supported 12
ordered, OPL keyword 64
overflow
and integer expressions 42
P
piecewise linear functions
pwlFunction 5
piecewise-linear functions 43
discontinuous 45
piecewise, OPL keyword 43
preprocessing
data 35
prev, OPL function 49
prevc, OPL function 49
processing order
preprocessing items 35
prod, OPL keyword 43
program unit, and local variables
(scripting) 101
properties
accessing value of 93
for IBM ILOG Script arrays 121
for strings 112
pushing and pulling data 23
pwlFunction, OPL keyword 5
Q
quotes 110, 117
R
R1C1 style not supported in Excel 29, 31
range float, data type 7
ranges 7
and set expressions 49
reading
from a database 26
from a spreadsheet 29
return, IBM ILOG Script keyword 102
rows
adding to a database 27
updating in a database 28
S
scheduling constraints in CP 63
scope hiding 68
scope of script variables 101
script functions
miscellaneous 132
script variables
declaration 101
inside a function definition 101
outside a function definition 101
reference to 93
scope 101
semi-colon 91
set expressions 49
and ranges 49
construction 49
functions 49
setof, OPL keyword 12
sets 12
allowed operations 12
and data consistency 32
and sparsity 21
generic 21
initializing 21
methods for sets 13
of tuples, initialization 16
ordered versus sorted 13
SheetConnection, OPL keyword 29
SheetRead, OPL keyword 29
shorthand operators 94
sign function 46
slicing
explicit/implicit 67
using key fields 10
slopes in piecewise-linear functions 43
sorted sets 13
sorted tuple sets 14
sparsity
and multidimensional arrays 8
and one-dimensional arrays 8
and sets 21
special numbers 104
special ordered sets, not supported 12
specialized constraints 63
spreadsheets
accessing named ranges 30
connecting to 29
data input/output 29
reading from 29
writing to 31
SQL requests
encryption 27
statements 99
conditional 99
in IBM ILOG Script for OPL 91
last 91
static, IBM ILOG Script keyword 102
stepFunction, OPL keyword 6
stepwise functions
stepFunction 6
stop function 132
string constraints 59
string, OPL keyword 5
strings 5, 110
automatic conversion to 111
functions 114
length 103
methods 112
operators 116
properties 112
structs in C, tuples in OPL 9
Index 137
sum, OPL keyword 43
symdiff, OPL keyword 49
syntax, in IBM ILOG Script 91
accessing property values 93
assignment operators 94
conditional statements 99
default values 103
expressions 93
function definition 102
identifiers 92
loops 99
other operators 97
reference to a script variable 93
shorthand 95
special keywords 96
special operators 96
variable declaration 101
T
then
See if-then-else 52
this, IBM ILOG Script keyword 96, 124,
125
time representation in IBM ILOG
Script 126
tuple indices in dexpr 11
tuple indices in labeled constraints 12
tuple patterns
in decision expressions 11
tuple sets
external initialization 16
referring to other sets with keys 33
sorted 13
tuple sets, sorted 14
tuple, OPL keyword 9
tuples 9
and data consistency 32
data types allowed, not allowed 11
duplicates in collections not
detected 11
initialization 20
keys in declaration 10
of parameters 66
filtering 67
tuples, limitations 11
U
undefined value 130
union, OPL keyword 49
updating a database 27
V
values
and functions (IBM ILOG Script) 119
values, in IBM ILOG Script
and functions
Booleans 117
dates 126
decimal numbers 104
hexadecimal numbers 104
numbers 104
objects 124
special numbers 104
values, in IBM ILOG Script (continued)
and functions (continued)
strings 110
assignment operators 94
default 103
null 130
of properties, accessing 93
undefined 130
values, in ILOG Script
and functions
octal numbers 104
var, IBM ILOG Script keyword 101
variables
See script variables
W
while, IBM ILOG Script keyword
loop syntax 99
with, IBM ILOG Script keyword 103
with, OPL keyword 32
138 OPL Language Reference Manual

Printed in USA

Navigation menu