The CVX Users' Guide User's Manual
User Manual:
Open the PDF directly: View PDF 
.
Page Count: 99
| Download | |
| Open PDF In Browser | View PDF |
The CVX Users’ Guide
Release 2.1
Michael C. Grant, Stephen P. Boyd
CVX Research, Inc.
June 10, 2015
CONTENTS
1
.
.
.
.
1
1
2
3
3
2
Installation
2.1 Supported platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 Installing a CVX Professional license . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3 Solvers included with CVX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
5
6
6
3
A quick start
3.1 Least squares . . . . . . . . . .
3.2 Bound-constrained least squares
3.3 Other norms and functions . . .
3.4 Other constraints . . . . . . . .
3.5 An optimal trade-off curve . . .
4
5
Introduction
1.1 What is CVX? . . . . . . . . . . . . . . .
1.2 What is disciplined convex programming?
1.3 What CVX is not . . . . . . . . . . . . . .
1.4 Licensing . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
9
9
11
12
14
15
The Basics
4.1 cvx_begin and cvx_end . . . .
4.2 Variables . . . . . . . . . . . . . .
4.3 Objective functions . . . . . . . . .
4.4 Constraints . . . . . . . . . . . . .
4.5 Functions . . . . . . . . . . . . . .
4.6 Set membership . . . . . . . . . .
4.7 Dual variables . . . . . . . . . . .
4.8 Assignment and expression holders
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
19
19
19
21
21
22
22
24
25
.
.
.
.
.
.
.
29
29
30
30
31
32
34
35
.
.
.
.
.
The DCP ruleset
5.1 A taxonomy of curvature . . . . . . . .
5.2 Top-level rules . . . . . . . . . . . . .
5.3 Constraints . . . . . . . . . . . . . . .
5.4 Expression rules . . . . . . . . . . . .
5.5 Functions . . . . . . . . . . . . . . . .
5.6 Compositions . . . . . . . . . . . . . .
5.7 Monotonicity in nonlinear compositions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
i
5.8
Scalar quadratic forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
6
Semidefinite programming mode
7
Geometric programming mode
41
7.1 Top-level rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
7.2 Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
7.3 Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
8
Solvers
8.1 Supported solvers . . . .
8.2 Selecting a solver . . . . .
8.3 Controlling screen output
8.4 Interpreting the results . .
8.5 Controlling precision . . .
8.6 Advanced solver settings .
9
Reference guide
9.1 Arithmetic operators
9.2 Built-in functions . .
9.3 New functions . . .
9.4 Sets . . . . . . . . .
9.5 Commands . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
10 Support
10.1 The CVX Forum . . . . .
10.2 Bug reports . . . . . . . .
10.3 What is a bug? . . . . . .
10.4 Handling numerical issues
10.5 CVX Professional support
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
39
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
45
45
46
46
47
48
50
.
.
.
.
.
53
53
54
55
60
61
.
.
.
.
.
63
63
63
64
65
66
11 Advanced topics
11.1 Eliminating quadratic forms . . . . . . .
11.2 Indexed dual variables . . . . . . . . . .
11.3 The successive approximation method . .
11.4 Power functions and p-norms . . . . . .
11.5 Overdetermined problems . . . . . . . .
11.6 Adding new functions to the atom library
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
67
67
68
69
71
72
73
12 License
12.1 CVX Professional License
12.2 CVX Standard License . .
12.3 The Free Solver Clause .
12.4 Bundled solvers . . . . .
12.5 Example library . . . . .
12.6 No Warranty . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
77
77
78
78
79
79
79
13 Citing CVX
ii
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
81
14 Credits and Acknowledgements
83
15 Using Gurobi with CVX
15.1 About Gurobi . . . . . . . . . . . . . . . . . . .
15.2 Using the bundled version of Gurobi . . . . . .
15.3 Using CVX with a standalone Gurobi installation
15.4 Selecting Gurobi as your default solver . . . . .
15.5 Obtaining support for CVX and Gurobi . . . . .
.
.
.
.
.
85
85
85
86
86
87
.
.
.
.
.
89
89
89
89
90
90
16 Using MOSEK with CVX
16.1 About MOSEK . . . . . . . . . . . . . . . . .
16.2 Using the bundled version of MOSEK . . . . .
16.3 Using CVX with separate MOSEK installation
16.4 Selecting MOSEK as your default solver . . .
16.5 Obtaining support for CVX and MOSEK . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Bibliography
91
Index
93
iii
iv
CHAPTER
ONE
INTRODUCTION
1.1 What is CVX?
CVX is a modeling system for constructing and solving disciplined convex programs (DCPs). CVX supports
a number of standard problem types, including linear and quadratic programs (LPs/QPs), second-order
cone programs (SOCPs), and semidefinite programs (SDPs). CVX can also solve much more complex
convex optimization problems, including many involving nondifferentiable functions, such as `1 norms. You
can use CVX to conveniently formulate and solve constrained norm minimization, entropy maximization,
determinant maximization, and many other convex programs. As of version 2.0, CVX also solves mixed
integer disciplined convex programs (MIDCPs) as well, with an appropriate integer-capable solver.
To use CVX effectively, you need to know at least a bit about convex optimization. For background on
convex optimization, see the book Convex Optimization [BV04] or the Stanford course EE364A.
CVX is implemented in Matlab, effectively turning Matlab into an optimization modeling language. Model
specifications are constructed using common Matlab operations and functions, and standard Matlab code
can be freely mixed with these specifications. This combination makes it simple to perform the calculations
needed to form optimization problems, or to process the results obtained from their solution. For example,
it is easy to compute an optimal trade-off curve by forming and solving a family of optimization problems
by varying the constraints. As another example, CVX can be used as a component of a larger system that
uses convex optimization, such as a branch and bound method, or an engineering design framework.
CVX provides special modes to simplify the construction of problems from two specific problem classes.
In semidefinite programming (SDP) mode, CVX applies a matrix interpretation to the inequality operator,
so that linear matrix inequalities (LMIs) and SDPs may be expressed in a more natural form. In geometric
programming (GP) mode, CVX accepts all of the special functions and combination rules of geometric programming, including monomials, posynomials, and generalized posynomials, and transforms such problems
into convex form so that they can be solved efficiently. For background on geometric programming, see this
tutorial paper [BKVH05].
Previous versions of CVX supported two free SQLP solvers, SeDuMi [Stu99] and SDPT3 [TTT03]. These
solvers are included with the CVX distribution. Starting with version 2.0, CVX supports two commercial
solvers as well, Gurobi and MOSEK. For more information, see Solvers.
The ability to use CVX with commercial solvers is a new capability that we have decided to include under
a new CVX Professional license model. Academic users will be able to utilize these features at no charge,
but commercial users will require a paid CVX Professional license. For more details, see Licensing.
1
The CVX Users’ Guide, Release 2.1
1.1.1 What’s new?
If you browse the source code and documentation, you will find indications of support for Octave with CVX.
However:
Note: Unfortunately, for average end users (this means you!), Octave will not work. The currently released
versions of Octave, including versions 3.8.0 and earlier, do not support CVX. Please do not waste your time
by trying!
We are working hard with the Octave team on final updates to bring CVX to Octave, and we anticipate
version 3.8.1 or 3.9.0 will be ready. We add this here to warn you not to interpret the mentions of Octave in
the code as a hidden code to try it yourself!
1.2 What is disciplined convex programming?
Disciplined convex programming is a methodology for constructing convex optimization problems proposed
by Michael Grant, Stephen Boyd, and Yinyu Ye [GBY06], [Gra04]. It is meant to support the formulation
and construction of optimization problems that the user intends from the outset to be convex.
Disciplined convex programming imposes a set of conventions or rules, which we call the DCP ruleset.
Problems which adhere to the ruleset can be rapidly and automatically verified as convex and converted to
solvable form. Problems that violate the ruleset are rejected—even when the problem is convex. That is not
to say that such problems cannot be solved using DCP; they just need to be rewritten in a way that conforms
to the DCP ruleset.
A detailed description of the DCP ruleset is given in The DCP ruleset. It is extremely important for anyone
who intends to actively use CVX to understand it. The ruleset is simple to learn, and is drawn from basic
principles of convex analysis. In return for accepting the restrictions imposed by the ruleset, we obtain
considerable benefits, such as automatic conversion of problems to solvable form, and full support for nondifferentiable functions. In practice, we have found that disciplined convex programs closely resemble their
natural mathematical forms.
1.2.1 Mixed integer problems
With version 2.0, CVX now supports mixed integer disciplined convex programs (MIDCPs). A MIDCP is
a model that obeys the same convexity rules as standard DCPs, except that one or more of its variables is
constrained to take on integral values. In other words, if the integer constraints are removed, the result is a
standard DCP.
Unlike a true DCP, a mixed integer problem is not convex. Finding the global optimum requires the combination of a traditional convex optimization algorithm with an exhaustive search such as a branch-and-bound
algorithm. Some CVX solvers do not include this second piece and therefore do not support MIDCPs; see
Solvers for more information. What is more, even the best solvers cannot guarantee that every moderatelysized MIDCP can be solved in a reasonable amount of time.
Mixed integer disciplined convex programming represents new territory for the CVX modeling framework—
and for the supporting solvers as well. While solvers for mixed integer linear and quadratic programs
2
Chapter 1. Introduction
The CVX Users’ Guide, Release 2.1
(MILP/MIQP) are reasonably mature, support for more general convex nonlinearities is a relatively new
development. We anticipate that MIDCP support will improve over time.
1.3 What CVX is not
CVX is not meant to be a tool for checking if your problem is convex. You need to know a bit about convex
optimization to effectively use CVX; otherwise you are the proverbial monkey at the typewriter, hoping to
(accidentally) type in a valid disciplined convex program. If you are not certain that your problem is convex
before you enter it into CVX, you are using the tool improperly, and your efforts will likely fail.
CVX is not meant for very large problems, so if your problem is very large (for example, a large image
processing or machine learning problem), CVX is unlikely to work well (or at all). For such problems you
will likely need to directly call a solver, or to develop your own methods, to get the efficiency you need.
For such problems CVX can play an important role, however. Before starting to develop a specialized
large-scale method, you can use CVX to solve scaled-down or simplified versions of the problem, to rapidly
experiment with exactly what problem you want to solve. For image reconstruction, for example, you might
use CVX to experiment with different problem formulations on 50 × 50 pixel images.
CVX will solve many medium and large scale problems, provided they have exploitable structure (such
as sparsity), and you avoid for loops, which can be slow in Matlab, and functions like log and exp
that require successive approximation. If you encounter difficulties in solving large problem instances,
consider posting your model to the CVX Forum; the CVX community may be able to suggest an equivalent
formulation that CVX can process more efficiently.
1.4 Licensing
CVX is free for use in both academic and commercial settings when paired with a free solver—including
the versions of SeDuMi and SDPT3 that are included with the package.
With version 2.0, we have added the ability to connect CVX to commercial solvers as well. This new
functionality is released under a CVX Professional product tier which we intend to license to commercial
users for a fee, and offer to academic users at no charge. The licensing structure is as follows:
• All users are free to use the standard features of CVX at no charge. This includes the ability to
construct and solve any of the models supported by the free solvers SeDuMi and SDPT3.
• Commercial users who wish to solve CVX models using Gurobi or MOSEK will need to purchase a
CVX Professional license. Please send an email to CVX Research for inquiries. for an availability
schedule and pricing details.
• Academic users may utilize the CVX Professional capability at no charge. To obtain an academic
license, please visit the Academic licenses page on the CVX Research web site.
The bulk of CVX remains open source under a slightly modified version of the GPL Version 2 license. A
small number of files that support the CVX Professional functionality remain closed source. If those files
are removed, the modified package remains fully functional using the free solvers, SeDuMi and SDPT3.
Users may freely modify, augment, and redistribute this free version of CVX, as long as all modifications
1.3. What CVX is not
3
The CVX Users’ Guide, Release 2.1
are themselves released under the same license. This includes adding support for new solvers released under
a free software license such as the GPL. For more details, please see the full Licensing section.
4
Chapter 1. Introduction
CHAPTER
TWO
INSTALLATION
2.1 Supported platforms
CVX is supported on 32-bit and 64-bit versions of Linux, Mac OSX, and Windows. For 32-bit platforms,
MATLAB version 7.5 (R2007b) or later is required; for 64-bit platforms, MATLAB version 7.8 (R2009a)
or later is required. There are some important platform-specific cautions, however:
• Gurobi support requires Matlab 7.7 (R2008b) or later.
• 32-bit Linux: the Gurobi solver is not available for this platform, as Gurobi is phasing out support for
32-bit Linux altogether.
• Older versions of Mac OS X (e.g. 10.5) ship with Java 1.5. The standard version of CVX works
properly on this platform, but CVX Professional support requires Java 1.6. To restore this support,
upgrade your operating system or Java installation.
As of version 2.0, support for versions 7.4 (R2007a) or older has been discontinued. If you need to use CVX
with these older versions of Matlab, please use CVX 1.22 or earlier, which will remain available indefinitely
on the CVX Research web site. However, this version is no longer supported, and will not receive bug
fixes or improvements. We strongly encourage you to update your Matlab installation to the latest version
possible.
Note: If you wish to use CVX with Gurobi or MOSEK, they must be installed and accessible from MATLAB before running cvx_setup. See below for more details.
1. Retrieve the latest version of CVX from the web site. You can download the package as either a .zip
file or a .tar.gz file.
2. Unpack the file anywhere you like; a directory called cvx will be created. There are two important
exceptions:
• Do not place CVX in Matlab’s own toolbox directory, Octave’s built-in scripts directory.
• Do not unpack a new version of CVX on top of an old one. We recommend moving the old
version out of the way, but do not delete it until you are sure the new version is working as you
expect.
3. Start Matlab or Octave. Do not add CVX to your path by hand.
5
The CVX Users’ Guide, Release 2.1
4. Change directories to the top of the CVX distribution, and run the cvx_setup command. For
example, if you installed CVX into C\personal\cvx on Windows, type these commands:
cd C:\personal\cvx
cvx_setup
at the MATLAB/Octave command prompt. If you installed CVX into ~/MATLAB/cvx on Linux or
a Mac, type these commands:
cd ~/MATLAB/cvx
cvx_setup
The cvx_setup function performs a variety of tasks to verify that your installation is correct, sets
your Matlab/Octave search path so it can find all of the CVX program files, and runs a simple test
problem to verify the installation.
5. In some cases—usually on Linux—the cvx_setup command may instruct you to create or modify
a startup.m file that allows you to use CVX without having to type cvx_setup every time you
re-start Matlab.
2.2 Installing a CVX Professional license
If you acquire a license key for CVX Professional, the only change required to the above steps is to include
the name of the license file as an input to the cvx_setup command. For example, if you saved your license
file to ~/licenses/cvx_license.mat on a Mac, this would be the modified command:
cd ~/MATLAB/cvx
cvx_setup ~/licenses/cvx_license.mat
If you have previously run cvx_setup without a license, or you need to replace your current license
with a new one, simply run cvx_setup again with the filename. Once the license has been accepted and
installed, you are free to move your license file anywhere you wish for safekeeping—CVX saves a copy in
its preferences.
2.3 Solvers included with CVX
All versions of CVX include copies of the solvers SeDuMi and SDPT3 in the directories cvx/sedumi and
cvx/sdpt3, respectively. When you run cvx_setup, CVX will automatically add these solvers to its solver
list.
If you have downloaded a CVX Professional Solver Bundle, then the solvers Gurobi and/or MOSEK will
be included with CVX as well. Use of these solvers requires a CVX Professional license. You may also
use your existing copies of these solvers with CVX as well. We have created special sections of this users’
guide for each solver:
• Gurobi: Using Gurobi with CVX
• MOSEK: Using MOSEK with CVX
6
Chapter 2. Installation
The CVX Users’ Guide, Release 2.1
For more general information on the solvers supported by CVX, an how to select a solver for your particular
problem, see the Solvers section.
2.3. Solvers included with CVX
7
The CVX Users’ Guide, Release 2.1
8
Chapter 2. Installation
CHAPTER
THREE
A QUICK START
Once you have installed CVX (see Installation), you can start using it by entering a CVX specification into a
Matlab script or function, or directly from the command prompt. To delineate CVX specifications from surrounding Matlab code, they are preceded with the statement cvx_begin and followed with the statement
cvx_end. A specification can include any ordinary Matlab statements, as well as special CVX-specific
commands for declaring primal and dual optimization variables and specifying constraints and objective
functions.
Within a CVX specification, optimization variables have no numerical value; instead, they are special Matlab
objects. This enables Matlab to distinguish between ordinary commands and CVX objective functions and
constraints. As CVX reads a problem specification, it builds an internal representation of the optimization
problem. If it encounters a violation of the rules of disciplined convex programming (such as an invalid use
of a composition rule or an invalid constraint), an error message is generated. When Matlab reaches the
cvx_end command, it completes the conversion of the CVX specification to a canonical form, and calls
the underlying core solver to solve it.
If the optimization is successful, the optimization variables declared in the CVX specification are converted
from objects to ordinary Matlab numerical values that can be used in any further Matlab calculations. In
addition, CVX also assigns a few other related Matlab variables. One, for example, gives the status of the
problem (i.e., whether an optimal solution was found, or the problem was determined to be infeasible or
unbounded). Another gives the optimal value of the problem. Dual variables can also be assigned.
This processing flow will become clearer as we introduce a number of simple examples. We invite the
reader to actually follow along with these examples in Matlab, by running the quickstart script found
in the examples subdirectory of the CVX distribution. For example, if you are on Windows, and you have
installed the CVX distribution in the directory D:\Matlab\cvx, then you would type
cd D:\Matlab\cvx\examples
quickstart
at the Matlab command prompt. The script will automatically print key excerpts of its code, and pause
periodically so you can examine its output. (Pressing “Enter” or “Return” resumes progress.)
3.1 Least squares
We first consider the most basic convex optimization problem, least-squares (also known as linear regression). In a least-squares problem, we seek x ∈ Rn that minimizes kAx − bk2 , where A ∈ Rm×n is skinny
and full rank (i.e., m ≥ n and Rank(A) = n). Let us create the data for a small test problem in Matlab:
9
The CVX Users’ Guide, Release 2.1
m = 16; n = 8;
A = randn(m,n);
b = randn(m,1);
Then the least-squares solution x = (AT A)−1 AT b is easily computed using the backslash operator:
x_ls = A \ b;
Using CVX, the same problem can be solved as follows:
cvx_begin
variable x(n)
minimize( norm(A*x-b) )
cvx_end
(The indentation is used for purely stylistic reasons and is optional.) Let us examine this specification line
by line:
• cvx_begin creates a placeholder for the new CVX specification, and prepares Matlab to accept
variable declarations, constraints, an objective function, and so forth.
• variable x(n) declares x to be an optimization variable of dimension n. CVX requires that all
problem variables be declared before they are used in the objective function or constraints.
• minimize( norm(A*x-b) ) specifies the objective function to be minimized.
• cvx_end signals the end of the CVX specification, and causes the problem to be solved.
Clearly there is no reason to use CVX to solve a simple least-squares problem. But this example serves
as sort of a “Hello world!” program in CVX; i.e., the simplest code segment that actually does something
useful.
When Matlab reaches the cvx_end command, the least-squares problem is solved, and the Matlab variable
x is overwritten with the solution of the least-squares problem, i.e., (AT A)−1 AT b. Now x is an ordinary
length-n numerical vector, identical to what would be obtained in the traditional approach, at least to within
the accuracy of the solver. In addition, several additional Matlab variables are created; for instance,
• cvx_optval contains the value of the objective function;
• cvx_status contains a string describing the status of the calculation (see Interpreting the results).
All of these quantities—x, cvx_optval, and cvx_status, etc.—may now be freely used in other
Matlab statements, just like any other numeric or string values. 1
There is not much room for error in specifying a simple least-squares problem, but if you make one, you
will get an error or warning message. For example, if you replace the objective function with
maximize( norm(A*x-b) );
which asks for the norm to be maximized, you will get an error message stating that a convex function
cannot be maximized (at least in disciplined convex programming):
1
If you type who or whos at the command prompt, you may see other, unfamiliar variables as well. Any variable that begins
with the prefix cvx_ is reserved for internal use by CVX itself, and should not be changed.
10
Chapter 3. A quick start
The CVX Users’ Guide, Release 2.1
??? Error using ==> maximize
Disciplined convex programming error:
Objective function in a maximization must be concave.
3.2 Bound-constrained least squares
Suppose we wish to add some simple upper and lower bounds to the least-squares problem above: i.e.,
minimize kAx − bk2
subject to l x u
where l and u are given data vectors with the same dimension as x. The vector inequality u v means
componentwise, i.e., ui ≤ vi for all i. We can no longer use the simple backslash notation to solve this
problem, but it can be transformed into a quadratic program (QP) which can be solved without difficulty
with a standard QP solver. 2
Let us provide some numeric values for l and u:
bnds = randn(n,2);
l = min( bnds, [], 2 );
u = max( bnds, [], 2 );
If you have the Matlab Optimization Toolbox, you can use quadprog to solve the problem as follows:
x_qp = quadprog( 2*A'*A, -2*A'*b, [], [], [], [], l, u );
This actually minimizes the square of the norm, which is the same as minimizing the norm itself. In contrast,
the CVX specification is given by
cvx_begin
variable x(n)
minimize( norm(A*x-b) )
subject to
l <= x <= u
cvx_end
Two new lines of CVX code have been added to the CVX specification:
• The subject to statement does nothing—CVX provides this statement simply to make specifications more readable. As with indentation, it is optional.
• The line l <= x <= u represents the 2n inequality constraints.
As before, when the cvx_end command is reached, the problem is solved, and the numerical solution is
assigned to the variable x. Incidentally, CVX will not transform this problem into a QP by squaring the
objective; instead, it will transform it into an SOCP. The result is the same, and the transformation is done
automatically.
In this example, as in our first, the CVX specification is longer than the Matlab alternative. On the other
hand, it is easier to read the CVX version and relate it to the original problem. In contrast, the quadprog
2
There are also a number of solvers specifically designed to solve bound-constrained least-squares problems, such as BCLS by
Michael Friedlander.
3.2. Bound-constrained least squares
11
The CVX Users’ Guide, Release 2.1
version requires us to know in advance the transformation to QP form, including the calculations such as
2*A’*A and -2*A’*b. For all but the simplest cases, a CVX specification is simpler, more readable, and
more compact than equivalent Matlab code to solve the same problem.
3.3 Other norms and functions
Now let us consider some alternatives to the least-squares problem. Norm minimization problems involving
the `∞ or `1 norms can be reformulated as LPs, and solved using a linear programming solver such as
linprog in the Matlab Optimization Toolbox; see, e.g., Section 6.1 of Convex Optimization. However,
because these norms are part of CVX’s base library of functions, CVX can handle these problems directly.
For example, to find the value of x that minimizes the Chebyshev norm kAx − bk∞ , we can employ the
linprog command from the Matlab Optimization Toolbox:
f
Ane
= [ zeros(n,1); 1
];
= [ +A,
-ones(m,1) ; ...
-A,
-ones(m,1) ];
bne = [ +b;
-b
];
xt
= linprog(f,Ane,bne);
x_cheb = xt(1:n,:);
With CVX, the same problem is specified as follows:
cvx_begin
variable x(n)
minimize( norm(A*x-b,Inf) )
cvx_end
The code based on linprog, and the CVX specification above will both solve the Chebyshev norm minimization problem, i.e., each will produce an x that minimizes kAx − bk∞ . Chebyshev norm minimization
problems can have multiple optimal points, however, so the particular x‘s produced by the two methods can
be different. The two points, however, must have the same value of kAx − bk∞ .
Similarly, to minimize the `1 norm k · k1 , we can use linprog as follows:
f
Aeq
lb
xzz
x_l1
=
=
=
=
=
[ zeros(n,1); ones(m,1); ones(m,1) ];
[ A,
-eye(m),
+eye(m)
];
[ -Inf(n,1); zeros(m,1); zeros(m,1) ];
linprog(f,[],[],Aeq,b,lb,[]);
xzz(1:n,:) - xzz(n+1:end,:);
The CVX version is, not surprisingly,
cvx_begin
variable x(n)
minimize( norm(A*x-b,1) )
cvx_end
CVX automatically transforms both of these problems into LPs, not unlike those generated manually for
linprog.
12
Chapter 3. A quick start
The CVX Users’ Guide, Release 2.1
The advantage that automatic transformation provides is magnified if we consider functions (and their resulting transformations) that are less well-known than the `∞ and `1 norms. For example, consider the
norm
kAx − bklgst,k = |Ax − b|[1] + · · · + |Ax − b|[k] ,
where |Ax − b|[i] denotes the ith largest element of the absolute values of the entries of Ax − b. This is
indeed a norm, albeit a fairly esoteric one. (When k = 1, it reduces to the `∞ norm; when k = m, the
dimension of Ax − b, it reduces to the `1 norm.) The problem of minimizing kAx − bklgst,k over x can
be cast as an LP, but the transformation is by no means obvious so we will omit it here. But this norm is
provided in the base CVX library, and has the name norm_largest, so to specify and solve the problem
using CVX is easy:
k = 5;
cvx_begin
variable x(n);
minimize( norm_largest(A*x-b,k) );
cvx_end
Unlike the `1 , `2 , or `∞ norms, this norm is not part of the standard Matlab distribution. Once you have
installed CVX, though, the norm is available as an ordinary Matlab function outside a CVX specification.
For example, once the code above is processed, x is a numerical vector, so we can type
cvx_optval
norm_largest(A*x-b,k)
The first line displays the optimal value as determined by CVX; the second recomputes the same value from
the optimal vector x as determined by CVX.
The list of supported nonlinear functions in CVX goes well beyond norm and norm_largest. For
example, consider the Huber penalty minimization problem
Pm
minimize
i=1 φ((Ax − b)i ) ,
with variable x ∈ Rn , where φ is the Huber penalty function
(
|z|2
|z| ≤ 1
φ(z) =
.
2|z| − 1 |z| ≥ 1
The Huber penalty function is convex, and has been provided in the CVX function library. So solving the
Huber penalty minimization problem in CVX is simple:
cvx_begin
variable x(n);
minimize( sum(huber(A*x-b)) );
cvx_end
CVX automatically transforms this problem into an SOCP, which the core solver then solves. (The CVX
user, however, does not need to know how the transformation is carried out.)
3.3. Other norms and functions
13
The CVX Users’ Guide, Release 2.1
3.4 Other constraints
We hope that, by now, it is not surprising that adding the simple bounds l x u to the problems above is
as simple as inserting the line l <= x <= u before the cvx_end statement in each CVX specification.
In fact, CVX supports more complex constraints as well. For example, let us define new matrices C and d
in Matlab as follows,
p = 4;
C = randn(p,n);
d = randn(p,1);
Now let us add an equality constraint and a nonlinear inequality constraint to the original least-squares
problem:
cvx_begin
variable x(n);
minimize( norm(A*x-b) );
subject to
C*x == d;
norm(x,Inf) <= 1;
cvx_end
Both of the added constraints conform to the DCP rules, and so are accepted by CVX. After the cvx_end
command, CVX converts this problem to an SOCP, and solves it.
Expressions using comparison operators (==, >=, etc.) behave quite differently when they involve CVX
optimization variables, or expressions constructed from CVX optimization variables, than when they involve
simple numeric values. For example, because x is a declared variable, the expression C*x==d causes a
constraint to be included in the CVX specification, and returns no value at all. On the other hand, outside of
a CVX specification, if x has an appropriate numeric value—for example immediately after the cvx_end
command—that same expression would return a vector of 1s and 0s, corresponding to the truth or falsity of
each equality. 3 Likewise, within a CVX specification, the statement norm(x,Inf)<=1 adds a nonlinear
constraint to the specification; outside of it, it returns a 1 or a 0 depending on the numeric value of x
(specifically, whether its `∞ -norm is less than or equal to, or more than, 1).
Because CVX is designed to support convex optimization, it must be able to verify that problems are convex.
To that end, CVX adopts certain rules that govern how constraint and objective expressions are constructed.
For example, CVX requires that the left- and right-hand sides of an equality constraint be affine. So a
constraint such as
norm(x,Inf) == 1;
results in the following error:
??? Error using ==> cvx.eq
Disciplined convex programming error:
Both sides of an equality constraint must be affine.
3
In fact, immediately after the cvx_end command above, you would likely find that most if not all of the values returned
would be 0. This is because, as is the case with many numerical algorithms, solutions are determined only to within some nonzero
numeric tolerance. So the equality constraints will be satisfied closely, but often not exactly.
14
Chapter 3. A quick start
The CVX Users’ Guide, Release 2.1
Inequality constraints of the form f (x) ≤ g(x) or g(x) ≥ f (x) are accepted only if f can be verified as
convex and g verified as concave. So a constraint such as
norm(x,Inf) >= 1;
results in the following error:
??? Error using ==> cvx.ge
Disciplined convex programming error:
The left-hand side of a ">=" inequality must be concave.
The specifics of the construction rules are discussed in more detail in The DCP ruleset. These rules are
relatively intuitive if you know the basics of convex analysis and convex optimization.
3.5 An optimal trade-off curve
For our final example in this section, let us show how traditional Matlab code and CVX specifications can
be mixed to form and solve multiple optimization problems. The following code solves the problem of
minimizing kAx − bk2 + γkxk1 , for a logarithmically spaced vector of (positive) values of γ. This gives us
points on the optimal trade-off curve between kAx − bk2 and kxk1 . An example of this curve is given in the
figure below.
gamma = logspace( -2, 2, 20 );
l2norm = zeros(size(gamma));
l1norm = zeros(size(gamma));
fprintf( 1, '
gamma
norm(x,1)
norm(A*x-b)\n' );
fprintf( 1, '---------------------------------------\n' );
for k = 1:length(gamma),
fprintf( 1, '%8.4e', gamma(k) );
cvx_begin
variable x(n);
minimize( norm(A*x-b)+gamma(k)*norm(x,1) );
cvx_end
l1norm(k) = norm(x,1);
l2norm(k) = norm(A*x-b);
fprintf( 1, '
%8.4e
%8.4e\n', l1norm(k), l2norm(k) );
end
plot( l1norm, l2norm );
xlabel( 'norm(x,1)' );
ylabel( 'norm(A*x-b)' );
grid on
The minimize statement above illustrates one of the construction rules to be discussed in The DCP ruleset.
A basic principle of convex analysis is that a convex function can be multiplied by a nonnegative scalar, or
added to another convex function, and the result is then convex. CVX recognizes such combinations and
allows them to be used anywhere a simple convex function can be—such as an objective function to be
minimized, or on the appropriate side of an inequality constraint. So in our example, the expression
norm(A*x-b)+gamma(k)*norm(x,1)
is recognized as convex by CVX, as long as gamma(k) is positive or zero. If gamma(k) were negative,
3.5. An optimal trade-off curve
15
The CVX Users’ Guide, Release 2.1
3.6
3.4
norm(A*x−b)
3.2
3
2.8
2.6
2.4
2.2
2
0
0.5
1
1.5
2
norm(x,1)
2.5
3
3.5
Figure 3.1: An example trade-off curve from the quickstart.m demo.
16
Chapter 3. A quick start
The CVX Users’ Guide, Release 2.1
then this expression becomes the sum of a convex term and a concave term, which causes CVX to generate
the following error:
??? Error using ==> cvx.plus
Disciplined convex programming error:
Addition of convex and concave terms is forbidden.
3.5. An optimal trade-off curve
17
The CVX Users’ Guide, Release 2.1
18
Chapter 3. A quick start
CHAPTER
FOUR
THE BASICS
4.1 cvx_begin and cvx_end
All CVX models must be preceded by the command cvx_begin and terminated with the command
cvx_end. All variable declarations, objective functions, and constraints should fall in between. The
cvx_begin command may include one more more modifiers:
cvx_begin quiet Prevents the model from producing any screen output while it is being solved.
cvx_begin sdp Invokes semidefinite programming mode.
cvx_begin gp Invokes geometric programming mode.
These modifiers may be combined when appropriate; for instance, cvx_begin sdp quiet invokes
SDP mode and silences the solver output.
4.2 Variables
All variables must be declared using the variable command (or variables command; see below)
before they can be used in constraints or an objective function. A variable command includes the name
of the variable, an optional dimension list, and one or more keywords that provide additional information
about the content or structure of the variable.
Variables can be real or complex scalars, vectors, matrices, or n-dimensional arrays. For instance,
variable X
variable Y(20,10)
variable Z(5,5,5)
declares a total of 326 (scalar) variables: a scalar X, a 20x10 matrix Y (containing 200 scalar variables), and
a 5x5x5 array Z (containing 125 scalar variables).
Variable declarations can also include one or more keywords to denote various structures or conditions on
the variable. For instance, to declare a complex variable, use the complex keyword:
variable w(50) complex
Nonnegative variables and symmetric/Hermitian positive semidefinite (PSD) matrices can be specified with
the nonnegative and semidefinite keywords, respectively:
19
The CVX Users’ Guide, Release 2.1
variable x(10) nonnegative
variable Z(5,5) semidefinite
variable Q(5,5) complex semidefinite
In this example, x is a nonnegative vector, and Z is a real symmetric PSD matrix and Q‘‘is a complex
Hermitian PSD matrix. As we will see below, ‘‘hermitian semidefinite
would be an equivalent choice for this third case.
For MIDCPs, the integer and binary keywords are used to declare integer and binary variables, respectively:
variable p(10) integer
variable q binary
A variety of keywords are available to help construct variables with matrix structure such as symmetry or
bandedness. For example, the code segment
variable Y(50,50) symmetric
variable Z(100,100) hermitian toeplitz
declares Y to be a real 50 × 50 symmetric matrix variable, and Z a 100 × 100 Hermitian Toeplitz matrix
variable. (Note that the hermitian keyword also specifies that the matrix is complex.) The currently
supported structure keywords are:
banded(lb,ub)
skew_symmetric
lower_bidiagonal
upper_bidiagonal
diagonal
symmetric
lower_hessenberg
upper_hankel
hankel
toeplitz
lower_triangular
upper_hessenberg
hermitian
tridiagonal
upper_triangular
The underscores can actually be omitted; so, for example, lower triangular is acceptable as well.
These keywords are self-explanatory with a couple of exceptions:
banded(lb,ub) the matrix is banded with a lower bandwidth lb and an upper bandwidth ub. If both lb
and ub are zero, then a diagonal matrix results. ub can be omitted, in which case it is set equal to lb.
For example, banded(1,1) (or banded(1)) is a tridiagonal matrix.
upper_hankel The matrix is Hankel (i.e., constant along antidiagonals), and zero below the central
antidiagonal, i.e., for i + j > n + 1.
When multiple keywords are supplied, the resulting matrix structure is determined by intersection. For
example, symmetric tridiagonal is a valid combination. That said, CVX does reject combinations
such as symmetric lower_triangular when a more reasonable alternative exists—diagonal, in
this case. Furthermore, if the keywords fully conflict, such that emph{no} non-zero matrix that satisfies all
keywords, an error will result.
Matrix-specific keywords can be applied to n-dimensional arrays as well: each 2-dimensional “slice” of the
array is given the stated structure. So for instance, the declaration
variable R(10,10,8) hermitian semidefinite
constructs 8 10 × 10 complex Hermitian PSD matrices, stored in the 2-D slices of R.
As flexible as the variable statement may be, it can only be used to declare a single variable, which
can be inconvenient if you have a lot of variables to declare. For this reason, the variables statement is
20
Chapter 4. The Basics
The CVX Users’ Guide, Release 2.1
provided which allows you to declare multiple variables; i.e.,
variables x1 x2 x3 y1(10) y2(10,10,10);
The one limitation of the variables command is that it cannot declare complex, integer, or structured
variables. These must be declared one at a time, using the singular variable command.
4.3 Objective functions
Declaring an objective function requires the use of the minimize or maximize function, as appropriate.
(For the benefit of our users whose English favors it, the synonyms minimise and maximise are provided
as well.) The objective function in a call to minimize must be convex; the objective function in a call to
maximize must be concave; for instance:
minimize( norm( x, 1 ) )
maximize( geo_mean( x ) )
At most one objective function may be declared in a CVX specification, and it must have a scalar value.
If no objective function is specified, the problem is interpreted as a feasibility problem, which is the same as
performing a minimization with the objective function set to zero. In this case, cvx_optval is either 0, if a
feasible point is found, or +Inf, if the constraints are not feasible.
4.4 Constraints
The following constraint types are supported in CVX:
• Equality == constraints, where both the left- and right-hand sides are affine expressions.
• Less-than <= inequality constraints, where the left-hand expression is convex, and the right-hand
expression is concave.
• Greater-than >= constraints, where the left-hand expression is concave, and the right-hand expression
is convex.
The non-equality operator ~= may never be used in a constraint; in any case, such constraints are rarely
convex. The latest version of CVX now allows you to chain inequalities together; e.g., l <= x <= u.
(Previous versions did not allow chained inequalities.)
Note the important distinction between the single equals =, which is an assignment, and the double equals
==, which denotes equality; for more on this distinction, see Assignment and expression holders below.
Strict inequalities < and > are accepted as well, but they are interpreted identically to their nonstrict counterparts. We strongly discourage their use, and a future version of CVX may remove them altogether. For
the reasoning behind this, please see the fuller discussion in Strict inequalities.
Inequality and equality constraints are applied in an elementwise fashion, matching the behavior of MATLAB itself. For instance, if A and B are m × n arrays, then A<=B is interpreted as mn (scalar) inequalities
A(i,j)<=B(i,j). When one side or the other is a scalar, that value is replicated; for instance, A>0 is
interpreted as A(i,j)>=0.
4.3. Objective functions
21
The CVX Users’ Guide, Release 2.1
The elementwise treatment of inequalities is altered in semidefinite programming mode; see that section for
more details.
CVX also supports a set membership constraint; see Set membership below.
4.5 Functions
The base CVX function library includes a variety of convex, concave, and affine functions which accept
CVX variables or expressions as arguments. Many are common Matlab functions such as sum, trace, diag,
sqrt, max, and min, re-implemented as needed to support CVX; others are new functions not found in Matlab.
A complete list of the functions in the base library can be found in Reference guide. It is also possible to
add your own new functions; see Adding new functions to the atom library.
An example of a function in the base library is the quadratic-over-linear function quad_over_lin:
(
xT x/y y > 0
f : Rn × R → R, f (x, y) =
+∞
y≤0
(The function also accepts complex x, but we’ll consider real x to keep things simple.) The quadratic-overlinear function is convex in x and y, and so can be used as an objective, in an appropriate constraint, or
in a more complicated expression. We can, for example, minimize the quadratic-over-linear function of
(Ax − b, cT x + d) using
minimize( quad_over_lin( A * x - b, c' * x + d ) );
inside a CVX specification, assuming x is a vector optimization variable, A is a matrix, b and c are vectors,
and d is a scalar. CVX recognizes this objective expression as a convex function, since it is the composition
of a convex function (the quadratic-over-linear function) with an affine function.
You can also use the function quad_over_lin outside a CVX specification. In this case, it just computes
its (numerical) value, given (numerical) arguments. If c’*x+d is positive, then the result is numerically
equivalent tp
( ( A * x - b )' * ( A * x - b ) ) / ( c' * x + d )
However, the quad_over_lin function also performs a domain check, so it returns Inf if c’*x+d is
zero or negative.
4.6 Set membership
CVX supports the definition and use of convex sets. The base library includes the cone of positive semidefinite n × n matrices, the second-order or Lorentz cone, and various norm balls. A complete list of sets
supplied in the base library is given in Sets.
Unfortunately, the Matlab language does not have a set membership operator, such as x in S, to denote
x ∈ S. So in CVX, we use a slightly different syntax to require that an expression is in a set. To represent a
set we use a function that returns an unnamed variable that is required to be in the set. Consider, for example,
Sn+ , the cone of symmetric positive semidefinite n × n matrices. In CVX, we represent this by the function
22
Chapter 4. The Basics
The CVX Users’ Guide, Release 2.1
semidefinite(n), which returns an unnamed new variable, that is constrained to be positive semidefinite. To
require that the matrix expression X be symmetric positive semidefinite, we use the syntax
X == semidefinite(n)
The literal meaning of this is that X is constrained to be equal to some unnamed variable, which is required
to be an n × n symmetric positive semidefinite matrix. This is, of course, equivalent to saying that X must
itself be symmetric positive semidefinite.
As an example, consider the constraint that a (matrix) variable X is a correlation matrix, i.e., it is symmetric,
has unit diagonal elements, and is positive semidefinite. In CVX we can declare such a variable and impose
these constraints using
variable X(n,n) symmetric;
X == semidefinite(n);
diag(X) == 1;
The second line here imposes the constraint that X be positive semidefinite. (You can read “==” here as “is”
or “is in”, so the second line can be read as X is positive semidefinite’.) The lefthand side of the third line is
a vector containing the diagonal elements of X, whose elements we require to be equal to one.
If this use of equality constraints to represent set membership remains confusing or simply aesthetically
displeasing, we have created a “pseudo-operator” that you can use in its place. So, for example, the
semidefinite constraint above can be replaced by
X semidefinite(n);
This is exactly equivalent to using the equality constraint operator, but if you find it more pleasing, feel free
to use it. Implementing this operator required some Matlab trickery, so don’t expect to be able to use it
outside of CVX models.
Sets can be combined in affine expressions, and we can constrain an affine expression to be in a convex set.
For example, we can impose a constraint of the form
A*X*A'-X B*semidefinite(n)*B';
where X is an n × n symmetric variable matrix, and A and B are n × n constant matrices. This constraint
requires that AXAT − X = BY B T , for some Y ∈ Sn+ .
CVX also supports sets whose elements are ordered lists of quantities. As an example, consider the secondorder or Lorentz cone,
Qm = { (x, y) ∈ Rm × R | kxk2 ≤ y } = epi k · k2 ,
where epi denotes the epigraph of a function. An element of Qm is an ordered list, with two elements: the
first is an m-vector, and the second is a scalar. We can use this cone to express the simple least-squares
problem from the section Least squares (in a fairly complicated way) as follows:
minimize y
subject to (Ax − b, y) ∈ Qm .
CVX uses Matlab’s cell array facility to mimic this notation:
4.6. Set membership
23
The CVX Users’ Guide, Release 2.1
cvx_begin
variables x(n) y;
minimize( y );
subject to
{ A*x-b, y } lorentz(m);
cvx_end
The function call lorentz(m) returns an unnamed variable (i.e., a pair consisting of a vector and a scalar
variable), constrained to lie in the Lorentz cone of length m. So the constraint in this specification requires
that the pair { A*x-b, y } lies in the appropriately-sized Lorentz cone.
4.7 Dual variables
When a disciplined convex program is solved, the associated dual problem is also solved. (In this context,
the original problem is called the primal problem.) The optimal dual variables, each of which is associated
with a constraint in the original problem, give valuable information about the original problem, such as the
sensitivities with respect to perturbing the constraints (c.f. Convex Optimization, chapter 5). To get access
to the optimal dual variables in CVX, you simply declare them, and associate them with the constraints.
Consider, for example, the LP
minimize cT x
subject to Ax b,
with variable x ∈ Rn , and m inequality constraints. To associate the dual variable y with the inequality
constraint Ax b in this LP, we use the following syntax:
n = size(A,2);
cvx_begin
variable x(n);
dual variable y;
minimize( c' * x );
subject to
y : A * x <= b;
cvx_end
The line
dual variable y
tells CVX that y will represent the dual variable, and the line
y : A * x <= b;
associates it with the inequality constraint. Notice how the colon : operator is being used in a different
manner than in standard Matlab, where it is used to construct numeric sequences like 1:10. This new
behavior is in effect only when a dual variable is present, so there should be no confusion or conflict. No
dimensions are given for y; they are automatically determined from the constraint with which it is associated.
For example, if m = 20, typing y at the Matlab command prompt immediately before cvx_end yields
24
Chapter 4. The Basics
The CVX Users’ Guide, Release 2.1
y =
cvx dual variable (20x1 vector)
It is not necessary to place the dual variable on the left side of the constraint; for example, the line above
can also be written in this way:
A * x <= b : y;
In addition, dual variables for inequality constraints will always be nonnegative, which means that the sense
of the inequality can be reversed without changing the dual variable’s value; i.e.,
b >= A * x : y;
yields an identical result. For equality constraints, on the other hand, swapping the left- and right- hand
sides of an equality constraint will negate the optimal value of the dual variable.
After the cvx_end statement is processed, and assuming the optimization was successful, CVX assigns
numerical values to x and y—the optimal primal and dual variable values, respectively. Optimal primal and
dual variables for this LP must satisfy the complementary slackness conditions
yi (b − Ax)i = 0,
i = 1, . . . , m.
You can check this in Matlab with the line
y .* (b-A*x)
which prints out the products of the entries of y and b-A*x, which should be nearly zero. This line must
be executed after the cvx_end command (which assigns numerical values to x and y); it will generate an
error if it is executed inside the CVX specification, where y and b-A*x are still just abstract expressions.
If the optimization is not successful, because either the problem is infeasible or unbounded, then x and y
will have different values. In the unbounded case, x will contain an unbounded direction; i.e., a point x
satisfying
cT x = −1,
Ax 0,
and y will be filled with NaN values, reflecting the fact that the dual problem is infeasible. In the infeasible
case, x is filled with NaN values, while y contains an unbounded dual direction; i.e., a point y satisfying
bT y = −1,
AT y = 0,
y0
Of course, the precise interpretation of primal and dual points and/or directions depends on the structure of
the problem. See references such as Convex Optimization for more on the interpretation of dual information.
CVX also supports the declaration of indexed dual variables. These prove useful when the number of constraints in a model (and, therefore, the number of dual variables) depends upon the parameters themselves.
For more information on indexed dual variables, see Indexed dual variables.
4.8 Assignment and expression holders
Anyone with experience with C or Matlab understands the difference between the single-equal assignment
operator = and the double-equal equality operator ==. This distinction is vitally important in CVX as
4.8. Assignment and expression holders
25
The CVX Users’ Guide, Release 2.1
well, and CVX takes steps to ensure that assignments are not used improperly. For instance, consider the
following code snippet:
variable X(n,n) symmetric;
X = semidefinite(n);
At first glance, the statement X = semidefinite(n); may look like it constrains X to be positive
semidefinite. But since the assignment operator is used, X is actually overwritten by the anonymous semidefinite variable instead. Fortunately, CVX forbids declared variables from being overwritten in this way; when
cvx_end is reached, this model would issue the following error:
??? Error using ==> cvx_end
The following cvx variable(s) have been overwritten:
X
This is often an indication that an equality constraint was
written with one equals '=' instead of two '=='. The model
must be rewritten before cvx can proceed.
We hope that this check will prevent at least some typographical errors from having frustrating consequences
in your models.
Despite this warning, assignments can be genuinely useful, so we encourage their use with appropriate care.
For instance, consider the following excerpt:
variables x y
z = 2 * x - y;
square( z ) <= 3;
quad_over_lin( x, z ) <= 1;
The construction z = 2 * x - y is not an equality constraint; it is an assignment. It is storing an
intermediate calculation 2 * x - y, which is an affine expression, which is then used later in two different
constraints. We call z an expression holder to differentiate it from a formally declared CVX variable.
Often it will be useful to accumulate an array of expressions into a single Matlab variable. Unfortunately,
a somewhat technical detail of the Matlab object model can cause problems in such cases. Consider this
construction:
variable u(9);
x(1) = 1;
for k = 1 : 9,
x(k+1) = sqrt( x(k) + u(k) );
end
This seems reasonable enough: x should be a vector whose first value is 1, and whose subsequent values are
concave CVX expressions. But if you try this in a CVX model, Matlab will give you a rather cryptic error:
??? The following error occurred converting from cvx to double:
Error using ==> double
Conversion to double from cvx is not possible.
The reason this occurs is that the Matlab variable x is initialized as a numeric array when the assignment
x(1)=1 is made; and Matlab will not permit CVX objects to be subsequently inserted into numeric arrays.
The solution is to explicitly declare x to be an expression holder before assigning values to it. We have
26
Chapter 4. The Basics
The CVX Users’ Guide, Release 2.1
provided keywords expression and expressions for just this purpose, for declaring a single or multiple expression holders for future assignment. Once an expression holder has been declared, you may freely insert
both numeric and CVX expressions into it. For example, the previous example can be corrected as follows:
variable u(9);
expression x(10);
x(1) = 1;
for k = 1 : 9,
x(k+1) = sqrt( x(k) + u(k) );
end
CVX will accept this construction without error. You can then use the concave expressions x(1), ..., x(10)
in any appropriate ways; for example, you could maximize x(10).
The differences between a variable object and an expression object are quite significant. A variable object
holds an optimization variable, and cannot be overwritten or assigned in the CVX specification. (After solving the problem, however, CVX will overwrite optimization variables with optimal values.) An expression
object, on the other hand, is initialized to zero, and should be thought of as a temporary place to store CVX
expressions; it can be assigned to, freely re-assigned, and overwritten in a CVX specification.
Of course, as our first example shows, it is not always necessary to declare an expression holder before it is
created or used. But doing so provides an extra measure of clarity to models, so we strongly recommend it.
4.8. Assignment and expression holders
27
The CVX Users’ Guide, Release 2.1
28
Chapter 4. The Basics
CHAPTER
FIVE
THE DCP RULESET
CVX enforces the conventions dictated by the disciplined convex programming ruleset, or DCP ruleset for
short. CVX will issue an error message whenever it encounters a violation of any of the rules, so it is
important to understand them before beginning to build models. The rules are drawn from basic principles
of convex analysis, and are easy to learn, once you’ve had an exposure to convex analysis and convex
optimization.
The DCP ruleset is a set of sufficient, but not necessary, conditions for convexity. So it is possible to
construct expressions
that violate the ruleset but are in fact convex. As an example consider the entropy
P
function, − ni=1 xi log xi , defined for x > 0, which is concave. If it is expressed as
- sum( x .* log( x ) )
CVX will reject it, because its concavity does not follow from any of the composition rules. (Specifically,
it violates the no-product rule described in Expression rules.) Problems involving entropy, however, can be
solved, by explicitly using the entropy function,
sum( entr( x ) )
which is in the base CVX library, and thus recognized as concave by CVX. If a convex (or concave) function
is not recognized as convex or concave by CVX, it can be added as a new atom; see Adding new functions
to the atom library.
√
As another example consider the function x2 + 1 = k[x 1]k2 , which is convex. If it is written as
norm( [ x 1 ] )
(assuming x is a scalar variable or affine expression) it will be recognized by CVX as a convex expression,
and therefore can be used in (appropriate) constraints and objectives. But if it is written as
sqrt( x^2 + 1 )
CVX will reject it, since convexity of this function does not follow from the CVX ruleset.
5.1 A taxonomy of curvature
In disciplined convex programming, a scalar expression is classified by its curvature. There are four categories of curvature: constant, affine, convex, and concave. For a function f : Rn → R defined on all Rn ,
29
The CVX Users’ Guide, Release 2.1
the categories have the following meanings:
constant
affine
convex
concave
f (αx + (1 − α)y) = f (x)
f (αx + (1 − α)y) = αf (x) + (1 − α)f (y)
f (αx + (1 − α)y) ≤ αf (x) + (1 − α)f (y)
f (αx + (1 − α)y) ≥ αf (x) + (1 − α)f (y)
∀x, y
∀x, y
∀x, y
∀x, y
∈ Rn ,
∈ Rn ,
∈ Rn ,
∈ Rn ,
α∈R
α∈R
α ∈ [0, 1]
α ∈ [0, 1]
Of course, there is significant overlap in these categories. For example, constant expressions are also affine,
and (real) affine expressions are both convex and concave.
Convex and concave expressions are real by definition. Complex constant and affine expressions can be
constructed, but their usage is more limited; for example, they cannot appear as the left- or right-hand side
of an inequality constraint.
5.2 Top-level rules
CVX supports three different types of disciplined convex programs:
• A minimization problem, consisting of a convex objective function and zero or more constraints.
• A maximization problem, consisting of a concave objective function and zero or more constraints.
• A feasibility problem, consisting of one or more constraints and no objective.
5.3 Constraints
Three types of constraints may be specified in disciplined convex programs:
• An equality constraint, constructed using ==, where both sides are affine.
• A less-than inequality constraint, using <=, where the left side is convex and the right side is concave.
• A greater-than inequality constraint, using >=, where the left side is concave and the right side is
convex.
Non-equality constraints, constructed using ~=, are never allowed. (Such constraints are not convex.)
One or both sides of an equality constraint may be complex; inequality constraints, on the other hand, must
be real. A complex equality constraint is equivalent to two real equality constraints, one for the real part
and one for the imaginary part. An equality constraint with a real side and a complex side has the effect of
constraining the imaginary part of the complex side to be zero.
As discussed in Set membership, CVX enforces set membership constraints (e.g., x ∈ S) using equality
constraints. The rule that both sides of an equality constraint must be affine applies to set membership
constraints as well. In fact, the returned value of set atoms like semidefinite() and lorentz()
is affine, so it is sufficient to simply verify the remaining portion of the set membership constraint. For
composite values like { x, y }, each element must be affine.
30
Chapter 5. The DCP ruleset
The CVX Users’ Guide, Release 2.1
5.3.1 Strict inequalities
As mentioned in Constraints, strict inequalities <, > are interpreted in an identical fashion to nonstrict
inequalities >=, <=. It is important to note that CVX cannot guarantee that an inequality will be strictly
satisfied at the solution it computes. This is not simply a choice we have made in CVX; it is a natural
consequence of both the underlying mathematics and the design of convex optimization solvers. For that
reason, we strongly discourage the use of strict inequalities in CVX, and a future version may remove them
altogether.
When a strict inequality is essential to your model, you may need to take additional steps to ensure compliance. In some cases, this can be accomplished through normalization. For instance, consider a set of
homogeneous equations and inequalities:
Ax = 0,
Cx 0,
x 0
Except for the strict inequality, x = 0 would be an acceptable solution; indeed the need to avoid the origin
is the very reason for the strict inequality. However, note that if a given x satisfies these constraints, then
so does αx for all α > 0. By eliminating this degree of freedom with normalization, we can eliminate the
strict inequality; for instance:
Ax = 0,
Cx 0,
x 0,
1T x = 1
If normalization is not a valid approach for your model, you may simply need to convert the strict inequality
into a non-strict one by adding a small offset; e.g., convert x > 0 to, say, x >= 1e-4. Note that the
bound needs to be large enough so that the underlying solver considers it numerically significant.
Finally, note that for some functions like log(x) and inv_pos(x), which have domains defined by
strict inequalities, the domain restriction is handled by the function itself. You do not need to add an explicit
constraint x > 0 to your model to guarantee that the solution is positive.
5.4 Expression rules
So far, the rules as stated are not particularly restrictive, in that all convex programs (disciplined or otherwise) typically adhere to them. What distinguishes disciplined convex programming from more general
convex programming are the rules governing the construction of the expressions used in objective functions
and constraints.
Disciplined convex programming determines the curvature of scalar expressions by recursively applying the
following rules. While this list may seem long, it is for the most part an enumeration of basic rules of convex
analysis for combining convex, concave, and affine forms: sums, multiplication by scalars, and so forth.
• A valid constant expression is
– any well-formed Matlab expression that evaluates to a finite value.
• A valid affine expression is
– a valid constant expression;
– a declared variable;
– a valid call to a function in the atom library with an affine result;
5.4. Expression rules
31
The CVX Users’ Guide, Release 2.1
– the sum or difference of affine expressions;
– the product of an affine expression and a constant.
• A valid convex expression is
– a valid constant or affine expression;
– a valid call to a function in the atom library with a convex result;
– an affine scalar raised to a constant power p ≥ 1, p 6= 3, 5, 7, 9, ...;
– a convex scalar quadratic form—see Scalar quadratic forms;
– the sum of two or more convex expressions;
– the difference between a convex expression and a concave expression;
– the product of a convex expression and a nonnegative constant;
– the product of a concave expression and a nonpositive constant;
– the negation of a concave expression.
• A valid concave expression is
– a valid constant or affine expression;
– a valid call to a function in the atom library with a concave result;
– a concave scalar raised to a power p ∈ (0, 1);
– a concave scalar quadratic form—see Scalar quadratic forms;
– the sum of two or more concave expressions;
– the difference between a concave expression and a convex expression;
– the product of a concave expression and a nonnegative constant;
– the product of a convex expression and a nonpositive constant;
– the negation of a convex expression.
If an expression cannot be categorized by this ruleset, it is rejected by CVX. For matrix and array expressions, these rules are applied on an elementwise basis. We note that the set of rules listed above is redundant;
there are much smaller, equivalent sets of rules.
Of particular note is that these expression rules generally forbid products between nonconstant expressions,
with the exception of scalar quadratic forms. For example, the expression x*sqrt(x) happens to be a
convex function of x, but its convexity cannot be verified using the CVX ruleset, and so is rejected. (It can
be expressed as pow_p(x,3/2), however.) We call this the no-product rule, and paying close attention to
it will go a long way to insuring that the expressions you construct are valid.
5.5 Functions
In CVX, functions are categorized in two attributes: curvature (constant, affine, convex, or concave) and
monotonicity (nondecreasing, nonincreasing, or nonmonotonic). Curvature determines the conditions under
32
Chapter 5. The DCP ruleset
The CVX Users’ Guide, Release 2.1
which they can appear in expressions according to the expression rules given above. Monotonicity determines how they can be used in function compositions, as we shall see in the next section.
For functions with only one argument, the categorization is straightforward. Some examples are given in
the table below.
Function
sum( x )
abs( x )
log( x )
sqrt( x )
Meaning
P
i xi
|x|
log x
√
x
Curvature
affine
convex
concave
concave
Monotonicity
nondecreasing
nonmonotonic
nondecreasing
nondecreasing
Following standard practice in convex analysis, convex functions are interpreted as +∞ when the argument
is outside the domain of the function, and concave functions are interpreted as −∞ when the argument is
outside its domain. In other words, convex and concave functions in CVX are interpreted as their extendedvalued extensions.
This has the effect of automatically constraining the argument of a function to be in the function’s domain.
For example, if we form sqrt(x+1) in a CVX specification, where x is a variable, then x will automatically be constrained to be larger than or equal to −1. There is no need to add a separate constraint, x>=-1,
to enforce this.
Monotonicity of a function is determined in the extended sense, i.e., including the values of the argument
outside its domain. For example, sqrt(x) is determined to be nondecreasing since its value is constant
(−∞) for negative values of its argument; then jumps up to 0 for argument zero, and increases for positive
values of its argument.
CVX does not consider a function to be convex or concave if it is so only over a portion of its domain, even
if the argument is constrained to lie in one of these portions. As an example, consider the function 1/x. This
function is convex for x > 0, and concave for x < 0. But you can never write 1/x in CVX (unless x is
constant), even if you have imposed a constraint such as x>=1, which restricts x to lie in the convex portion
of function 1/x. You can use the CVX function inv_pos(x), defined as 1/x for x > 0 and ∞ otherwise,
for the convex portion of 1/x; CVX recognizes this function as convex and nonincreasing. In CVX, you can
express the concave portion of 1/x, where x is negative, using -inv_pos(-x), which will be correctly
recognized as concave and nonincreasing.
For functions with multiple arguments, curvature is always considered jointly, but monotonicity can be
considered on an argument-by-argument basis. For example, the function quad_over_lin(x,y)
(
|x|2 /y y > 0
fquad_over_lin (x, y) =
+∞
y≤0
is jointly convex in both x and y, but it is monotonic (nonincreasing) only in y.
Some functions are convex, concave, or affine only for a subset of its arguments. For example, the function
norm(x,p) where p \geq 1 is convex only in its first argument. Whenever this function is used in a
CVX specification, then, the remaining arguments must be constant, or CVX will issue an error message.
Such arguments correspond to a function’s parameters in mathematical terminology; e.g.,
fp (x) : Rn → R,
fp (x) , kxkp
So it seems fitting that we should refer to such arguments as parameters in this context as well. Henceforth, whenever we speak of a CVX function as being convex, concave, or affine, we will assume that its
parameters are known and have been given appropriate, constant values.
5.5. Functions
33
The CVX Users’ Guide, Release 2.1
5.6 Compositions
A basic rule of convex analysis is that convexity is closed under composition with an affine mapping. This
is part of the DCP ruleset as well:
• A convex, concave, or affine function may accept an affine expression (of compatible size) as an
argument. The result is convex, concave, or affine, respectively.
For example, consider the function square(x), which is provided in the CVX atom library. This function
squares its argument; i.e., it computes x.*x. (For array arguments, it squares each element independently.)
It is in the CVX atom library, and known to be convex, provided its argument is real. So if x is a real variable
of dimension n, a is a constant n-vector, and b is a constant, the expression
square( a' * x + b )
is accepted by CVX, which knows that it is convex.
The affine composition rule above is a special case of a more sophisticated composition rule, which we describe now. We consider a function, of known curvature and monotonicity, that accepts multiple arguments.
For convex functions, the rules are:
• If the function is nondecreasing in an argument, that argument must be convex.
• If the function is nonincreasing in an argument, that argument must be concave.
• If the function is neither nondecreasing or nonincreasing in an argument, that argument must be affine.
If each argument of the function satisfies these rules, then the expression is accepted by CVX, and is classified as convex. Recall that a constant or affine expression is both convex and concave, so any argument can
be affine, including as a special case, constant.
The corresponding rules for a concave function are as follows:
• If the function is nondecreasing in an argument, that argument must be concave.
• If the function is nonincreasing in an argument, that argument must be convex.
• If the function is neither nondecreasing or nonincreasing in an argument, that argument must be affine.
In this case, the expression is accepted by CVX, and classified as concave.
For more background on these composition rules, see Convex Optimization, Section 3.2.4. In fact, with the
exception of scalar quadratic expressions, the entire DCP ruleset can be thought of as special cases of these
six rules.
Let us examine some examples. The maximum function is convex and nondecreasing in every argument, so
it can accept any convex expressions as arguments. For example, if x is a vector variable, then
max( abs( x ) )
obeys the first of the six composition rules and is therefore accepted by CVX, and classified as convex.
As another example, consider the sum function, which is both convex and concave (since it is affine), and
nondecreasing in each argument. Therefore the expressions
sum( square( x ) )
sum( sqrt( x ) )
34
Chapter 5. The DCP ruleset
The CVX Users’ Guide, Release 2.1
are recognized as valid in CVX, and classified as convex and concave, respectively. The first one follows
from the first rule for convex functions; and the second one follows from the first rule for concave functions.
Most people who know basic convex analysis like to think of these examples in terms of the more specific
rules: a maximum of convex functions is convex, and a sum of convex (concave) functions is convex (concave). But these rules are just special cases of the general composition rules above. Some other well known
basic rules that follow from the general composition rules are:
• a nonnegative multiple of a convex (concave) function is convex (concave);
• a nonpositive multiple of a convex (concave) function is concave (convex).
Now we consider a more complex example in depth. Suppose x is a vector variable, and A, b, and f are
constants with appropriate dimensions. CVX recognizes the expression
sqrt(f'*x) + min(4,1.3-norm(A*x-b))
as concave.
Consider the term sqrt(f’*x).
CVX recognizes that sqrt is concave and
f’*x is affine, so it concludes that sqrt(f’*x) is concave. Now consider the second term
min(4,1.3-norm(A*x-b)). CVX recognizes that min is concave and nondecreasing, so it can accept concave arguments. CVX recognizes that 1.3-norm(A*x-b) is concave, since it is the difference
of a constant and a convex function. So CVX concludes that the second term is also concave. The whole
expression is then recognized as concave, since it is the sum of two concave functions.
The composition rules are sufficient but not necessary for the classification to be correct, so some expressions
which are in fact convex or concave will fail to satisfy them, and so will be rejected by CVX. For example,
if x is a vector variable, the expression
sqrt( sum( square( x ) ) )
is rejected by CVX, because there is no rule governing the composition of a concave nondecreasing function
with a convex function. Of course, the workaround is simple in this case: use norm( x ) instead, since
norm is in the atom library and known by CVX to be convex.
5.7 Monotonicity in nonlinear compositions
Monotonicity is a critical aspect of the rules for nonlinear compositions. This has some consequences that
are not so obvious, as we shall demonstrate here by example. Consider the expression
square( square( x ) + 1 )
where x is a scalar variable. This expression is in fact convex, since (x2 + 1)2 = x4 + 2x2 + 1 is convex.
But CVX will reject the expression, because the outer square cannot accept a convex argument. Indeed,
the square of a convex function is not, in general, convex: for example, (x2 − 1)2 = x4 − 2x2 + 1 is not
convex.
There are several ways to modify the expression above to comply with the ruleset. One way is to write it as
x^4 + 2*x^2 + 1, which CVX recognizes as convex, since CVX allows positive even integer powers
using the ^ operator. (Note that the same technique, applied to the function (x2 − 1)2 , will fail, since its
second term is concave.)
5.7. Monotonicity in nonlinear compositions
35
The CVX Users’ Guide, Release 2.1
Another approach is to use the alternate outer function square_pos, included in the CVX library, which
represents the function (x+ )2 , where x+ = max{0, x}. Obviously, square and square_pos coincide
when their arguments are nonnegative. But square_pos is nondecreasing, so it can accept a convex
argument. Thus, the expression
square_pos( square( x ) + 1 )
is mathematically equivalent to the rejected version above (since the argument to the outer function is always
positive), but it satisfies the DCP ruleset and is therefore accepted by CVX.
This is the reason several functions in the CVX atom library come in two forms: the “natural” form, and one
that is modified in such a way that it is monotonic, and can therefore be used in compositions. Other such
“monotonic extensions” include sum_square_pos and quad_pos_over_lin. If you are implementing a new function yourself, you might wish to consider if a monotonic extension of that function would
also be useful.
5.8 Scalar quadratic forms
In its pure form, the DCP ruleset forbids even the use of simple quadratic expressions such as x * x
(assuming x is a scalar variable). For practical reasons, we have chosen to make an exception to the ruleset
to allow for the recognition of certain specific quadratic forms that map directly to certain convex quadratic
functions (or their concave negatives) in the CVX atom library:
x .* x
conj( x ) .* x
y’ * y
(A*x-b)’*Q*(Ax-b)
square( x ) (real x)
square_abs( x )
sum_square_abs( y )
quad_form( A*x - b, Q )
CVX detects the quadratic expressions such as those on the left above, and determines whether or not they
are convex or concave; and if so, translates them to an equivalent function call, such as those on the right
above.
CVX examines each single product of affine expressions, and each single squaring of an affine expression,
checking for convexity; it will not check, for example, sums of products of affine expressions. For example,
given scalar variables x and y, the expression
x ^ 2 + 2 * x * y + y ^2
will cause an error in CVX, because the second of the three terms 2 * x * y, is neither convex nor
concave. But the equivalent expressions
( x + y ) ^ 2
( x + y ) * ( x + y )
will be accepted.
CVX actually completes the square when it comes across a scalar quadratic form, so the form need not be
symmetric. For example, if z is a vector variable, a, b are constants, and Q is positive definite, then
( z + a )' * Q * ( z + b )
36
Chapter 5. The DCP ruleset
The CVX Users’ Guide, Release 2.1
will be recognized as convex. Once a quadratic form has been verified by CVX, it can be freely used in any
way that a normal convex or concave expression can be, as described in Expression rules.
Quadratic forms should actually be used less frequently in disciplined convex programming than in a more
traditional mathematical programming framework, where a quadratic form is often a smooth substitute for
a nonsmooth form that one truly wishes to use. In CVX, such substitutions are rarely necessary, because of
its support for nonsmooth functions. For example, the constraint
sum( ( A * x - b ) .^ 2 ) <= 1
is equivalently represented using the Euclidean norm:
norm( A * x - b ) <= 1
With modern solvers, the second form is more naturally represented using a second-order cone constraint—
so the second form may actually be more efficient. In fact, in our experience, the non-squared form will
often be handled more accurately. So we strongly encourage you to re-evaluate the use of quadratic forms
in your models, in light of the new capabilities afforded by disciplined convex programming.
5.8. Scalar quadratic forms
37
The CVX Users’ Guide, Release 2.1
38
Chapter 5. The DCP ruleset
CHAPTER
SIX
SEMIDEFINITE PROGRAMMING MODE
Those who are familiar with semidefinite programming (SDP) know that the constraints that utilize the set
semidefinite(n) in the discussion on Set membership above are, in practice, typically expressed using
linear matrix inequality (LMI) notation. For example, given X = X T ∈ Rn×n , the constraint X 0
denotes that X ∈ Sn+ ; that is, that X is positive semidefinite.
CVX provides a special SDP mode that allows this LMI notation to be employed inside CVX models using
Matlab’s standard inequality operators >=, <=. In order to use it, one simply begins a model with the
statement cvx_begin sdp or cvx_begin SDP instead of simply cvx_begin.
When SDP mode is engaged, CVX interprets certain inequality constraints in a different manner. To be
specific:
• Equality constraints are interpreted the same (i.e., elementwise).
• Inequality constraints involving vectors and scalars are interpreted the same; i.e., elementwise.
• Inequality constraints involving non-square matrices are disallowed; attempting to use them causes
an error. If you wish to do true elementwise comparison of matrices X and Y, use a vectorization
operation X(:) <= Y(:) or vec( X ) <= vec( Y ). (vec is a function provided by CVX
that is equivalent to the colon operation.)
• Inequality constraints involving real, square matrices are interpreted as follows:
X >= Y
X <= Y
becomes
becomes
X - Y == semidefinite(n)
Y - X == semidefinite(n)
If either side is complex, then the inequalities are interpreted as follows:
X >= Y
X <= Y
becomes
becomes
X - Y == hermitian_semidefinite(n)
Y - X == hermitian_semidefinite(n)
• There is one additional restriction: both X and Y must be the same size, or one must be the scalar zero.
For example, if X and Y are matrices of size n,
X >= 1
X >= ones(n,n)
X >= 0
or
or
or
1 >= Y
ones(n,n) >= Y
0 >= Y
illegal
legal
legal
In effect, CVX enforces a stricter interpretation of the inequality operators for LMI constraints.
• Note that LMI constraints enforce symmetry (real or Hermitian, as appropriate) on their inputs. Unlike
SDPSOL, CVX does not extract the symmetric part for you: you must take care to insure symmetry
39
The CVX Users’ Guide, Release 2.1
yourself. Since CVX supports the declaration of symmetric matrices, this is reasonably straightforward. If CVX cannot determine that an LMI is symmetric to within a reasonable numeric tolerance,
a warning will be issued. We have provided a function sym(X) that extracts the symmetric part of a
square matrix; that is, sym(X) = 0.5*(X+X’).
• A dual variable, if supplied, will be applied to the converted equality constraint. It will be given a
positive semidefinite value if an optimal point is found.
So, for example, the CVX model found in the file examples/closest_toeplitz_sdp.m,
cvx_begin
variable Z(n,n) hermitian toeplitz
dual variable Q
minimize( norm( Z - P, 'fro' ) )
Z == hermitian_semidefinite( n ) : Q;
cvx_end
can also be written as follows:
cvx_begin sdp
variable Z(n,n) hermitian toeplitz
dual variable Q
minimize( norm( Z - P, 'fro' ) )
Z >= 0 : Q;
cvx_end
Many other examples in the CVX example library utilize semidefinite constraints; and all of them use SDP
mode. To find them, simply search for the text cvx_begin sdp in the examples/ subdirectory tree
using your favorite file search tool. One of these examples is reproduced in Indexed dual variables.
Since semidefinite programming is popular, some may wonder why SDP mode is not the default behavior.
The reason for this is that we place a strong emphasis on maintaining consistency between Matlab’s native
behavior and that of CVX. Using the >=, <=, >, < operators to create LMIs represents a deviation from
that ideal. For example, the expression Z >= 0 in the example above constrains the variable Z to be
positive semidefinite. But after the model has been solved and Z has been replaced with a numeric value,
the expression Z >= 0 will test for the elementwise nonnegativity of Z. To verify that the numeric value of
Z is, in fact, positive semidefinite, you must perform a test like min(eig(Z)) >= 0.
40
Chapter 6. Semidefinite programming mode
CHAPTER
SEVEN
GEOMETRIC PROGRAMMING MODE
Geometric programs (GPs) are special mathematical programs that can be converted to convex form using
a change of variables. The convex form of GPs can be expressed as DCPs, but CVX also provides a special
mode that allows a GP to be specified in its native form. CVX will automatically perform the necessary
conversion, compute a numerical solution, and translate the results back to the original problem.
To utilize GP mode, you must begin your CVX specification with the command cvx_begin gp or
cvx_begin GP instead of simply cvx_begin. For example, the following code, found in the example library at gp/max_volume_box.m, determines the maximum volume box subject to various area
and ratio constraints:
cvx_begin gp
variables w h d
maximize( w * h * d )
subject to
2*(h*w+h*d) <= Awall;
w*d <= Afloor;
alpha <= h/w >= beta;
gamma <= d/w <= delta;
cvx_end
As the example illustrates, CVX supports the construction of monomials and posynomials using addition,
multiplication, division (when appropriate), and powers. In addition, CVX supports the construction of generalized geometric programs (GGPs), by permitting the use of generalized posynomials wherever posynomials are permitted in standard GP. More information about generalized geometric programs is provided in
this tutorial.
The solvers used in this version of CVX do not support geometric programming natively. Instead, they are
solved using the successive approximation technique described in The successive approximation method.
This means that solving GPs can be slow, but for small and medium sized problems, the method works well.
In the remainder of this section, we will describe specific rules that apply when constructing models in GP
mode.
7.1 Top-level rules
CVX supports three types of geometric programs:
41
The CVX Users’ Guide, Release 2.1
• A minimization problem, consisting of a generalized posynomial objective and zero or more constraints.
• A maximization problem, consisting of a monomial objective and zero or more constraints.
• A feasibility problem, consisting of one or more constraints.
The asymmetry between minimizations and maximizations—specifically, that only monomial objectives are
allowed in the latter—is an unavoidable artifact of the geometry of GPs and GGPs.
7.2 Constraints
Three types of constraints may be specified in geometric programs:
• An equality constraint, constructed using ==, where both sides are monomials.
• A less-than inequality constraint <= where the left side is a generalized posynomial and the right side
is a monomial.
• A greater-than inequality constraint >= where the left side is a monomial and the right side is a
generalized posynomial.
As with DCPs, non-equality constraints are not permitted; and while strict inequalities <, > are supported,
they are treated as non-strict inequalities and should therefore be avoided.
7.3 Expressions
The basic building blocks of generalized geometric programming are monomials, posynomials, and generalized posynomials. A valid monomial is
• a declared variable;
• the product of two or more monomials;
• the ratio of two monomials;
• a monomial raised to a real power; or
• a call to one of the following functions with monomial arguments: prod, cumprod, geo_mean,
sqrt.
A valid posynomial expression is
• a valid monomial;
• the sum of two or more posynomials;
• the product of two or more posynomials;
• the ratio of a posynomial and a monomial;
• a posynomial raised to a positive integral power; or
42
Chapter 7. Geometric programming mode
The CVX Users’ Guide, Release 2.1
• a call to one of the following functions with posynomial arguments: sum, cumsum, mean, prod,
cumprod.
A valid generalized posynomial expression is
• a valid posynomial;
• the sum of two or more generalized posynomials;
• the product of two or more generalized posynomials;
• the ratio of a generalized posynomial and a monomial;
• a generalized posynomial raised to a positive real power; or
• a call to one of the following functions with arguments that are generalized posynomials: sum,
cumsum, mean, prod, cumprod, geo_mean, sqrt, norm, sum_largest, norm_largest.
It is entirely possible to create and manipulate arrays of monomials, posynomials, and/or generalized posynomials in CVX, in which case these rules extend in an obvious manner. For example, the product of two
monomial matrices produces a matrix whose entries are polynomials (or monomials in special cases).
7.3. Expressions
43
The CVX Users’ Guide, Release 2.1
44
Chapter 7. Geometric programming mode
CHAPTER
EIGHT
SOLVERS
8.1 Supported solvers
This version of CVX supports four solvers, each with different capabilities:
Solver name
SeDuMi
SDPT3
Gurobi
MOSEK
GLPK
LP
Y
Y
Y
Y
Y
QP
Y
Y
Y
Y
N
SOCP
Y
Y
Y
Y
N
SDP
Y
Y
N
Y*
N
GP
E
E
N
E
N
Integer
N
N
Y
Y
Y
MATLAB
Y
Y
P
P
N
Octave
soon
soon
soon
soon
soon
(key: Y = Yes, N = No, E = Experimental, P = CVX Professional license required, * = Mosek 7 or later is
required.)
Each solver has different capabilities and different levels of performance. For instance, SeDuMi [Stu99],
SDPT3 [TTT03], and MOSEK 7 support all of the continuous (non-integer) models that CVX itself supports,
while Gurobi is more limited, in that it does not support semidefinite constraints; and GLPK is limited even
further. On the other hand, Gurobi, GLPK, and MOSEK support integer consraints, while SeDuMi and
SDPT3 do not.
SeDuMi and SDPT3 are included with the standard CVX distribution, so you do not need to download an
additional solver to start using CVX. We have also entered into contractual arrangements with the developers
of Gurobi and MOSEK that allow us to ship their binaries with CVX as well, but using those solvers
requires a CVX Professional license. Due to license differences, we are not able to supply GLPK with
CVX. However,
If you are having difficulty with one solver, please try another. No one solver performs better than the others
on every model CVX can generate—including commercial solvers. That said, if you encounter a problem
that one solver can handle well and another cannot, please send us a bug report (see Support) and we will
forward the results to the solver’s authors.
We have created special sections in this user guide for using Gurobi and MOSEK with CVX:
• Gurobi: Using Gurobi with CVX
• Mosek: Using MOSEK with CVX
Support for GLPK should be considered experimental, and has been provided primarly to support upcoming
Octave capability (that is not ready yet.)
45
The CVX Users’ Guide, Release 2.1
8.2 Selecting a solver
The default solver is currently SDPT3. We have found that SeDuMi is faster for most problems, but unfortunately not as reliable. None of the solvers are perfect, however, and you may find for your application that
another solver is preferred.
To see which solver is currently selected, simply type
cvx_solver
To change the current solver, simply follow the cvx_solver with the name of your chosen solver. For
example, to select SeDuMi, type
cvx_solver sedumi
The cvx_solver command is case insensitive, so cvx_solver SeDuMi will work just fine as well.
If you issue this command inside a model—that is, between cvx_begin and cvx_end it will change the
solver only for that model; the next model will use the previous choice. If, only the other hand, you issue a
cvx_solver command outside of a model, it will change the solver used for the remainder of your Matlab
session (or until you change it again).
If you would like to change the default solver permanently—that is, so that it remains the default even if you
quit and re-start Matlab—then make sure it is set properly, and then issue the command
cvx_save_prefs
This command saves not only your solver choice, but also your settings for cvx_expert,
cvx_power_warning, and cvx_precision as well.
8.3 Controlling screen output
Once you gain confidence in using CVX and start incorporating it into your larger algorithms and programs,
you are likely going to want to silence the messages it delivers to the screen. To do so, simply add the
quiet keyword to the cvx_begin command; that is,
cvx_begin quiet
...
cvx_end
Previous versions of CVX utilized a separate cvx_quiet command and that command is still available
in this version as well, if you prefer it. Entering cvx_quiet true suppresses screen output from the
solver, while entering cvx_quiet false restores the screen output. If you enter these commands within
a model—that is, between cvx_begin and cvx_end—it will affect only that model. If you enter it
outside of a model, it will affect all subsequent models. Entering cvx_quiet with no arguments returns the
current setting.
46
Chapter 8. Solvers
The CVX Users’ Guide, Release 2.1
8.4 Interpreting the results
After a complete CVX specification has been entered and the cvx_end command issued, the solver is called
to generate a numerical result. It proceeds to replace the variables in your model with the computed numerical values, and creates the variable cvx_optval containing the value of the objective function. It also
summarizes the result of its efforts in the form of a string named cvx_status. The possible values of
cvx_status are as follows:
Solved A complementary (primal and dual) solution has been found. The primal and dual variables are
replaced with their computed values, and the the optimal value of the problem is placed in cvx_optval
(which, by convention, is 0 for feasibility problems).
Unbounded The solver has determined that the problem is unbounded. The value of cvx_optval is
set to -Inf for minimizations, and +Inf for maximizations. (Feasibility problems, by construction,
never produce an Unbounded status.) The values of any dual variables are replaced with NaN, as
the dual problem is in fact infeasible.
For unbounded problems, CVX stores an unbounded direction into the problem variables. This is is
a direction along which the feasible set is unbounded, and the optimal value approaches ±∞. It is
important to understand that this value is very likely not a feasible point. If a feasible point is required,
the problem should be re-solved as a feasibility problem by omitting the objective. Mathematically
speaking, given an unbounded direction v and a feasible point x, x + tv is feasible for all t ≥ 0, and
the objective tends to −∞ (for minimizations; +∞ for maximizations) as t → +∞ itself.
Infeasible The problem has been proven to be infeasible through the discovery of an unbounded direction. The values of the variables are filled with NaN, and the value of cvx_optval is set to +Inf
for minimizations and feasibility problems, and -Inf for maximizations.
Associated with a provably infeasible problem is an unbounded dual direction. Appropriate components of this direction are stored in the dual variables. Similarly to the Unbounded case, it is
important to understand that the unbounded dual direction is very likely not a feasible dual point.
Inaccurate/Solved, Inaccurate/Unbounded, Inaccurate/Infeasible These three status values indicate that the solver was unable to make a determination to within the default numerical
tolerance. However, it determined that the results obtained satisfied a “relaxed” tolerance leve and
therefore may still be suitable for further use. If this occurs, you should test the validity of the computed solution before using it in further calculations. See Controlling precision for a more advanced
discussion of solver tolerances and how to make adjustments.
Suboptimal This status is possible only for mixed-integer problems. It is returned when the branching
algorithm has discovered at least one feasible integer solution, but it was unable to continue the search
process to global optimality. This will occur if the solver is required to terminate due to a time limit
or a forced interruption (for example, if the user types Ctrl-C.)
Failed The solver failed to make sufficient progress towards a solution, even to within the “relaxed” tolerance setting. The values of cvx_optval and primal and dual variables are filled with NaN. This result
can occur because of numerical problems within SeDuMi, often because the problem is particularly
“nasty” in some way (e.g., a non-zero duality gap).
Overdetermined The presolver has determined that the problem has more equality constraints than
variables, which means that the coefficient matrix of the equality constraints is singular. In practice,
such problems are often, but not always, infeasible. Unfortunately, solvers typically cannot handle
8.4. Interpreting the results
47
The CVX Users’ Guide, Release 2.1
such problems, so a precise conclusion cannot be reached. The situations that most commonly produce
an Overdetermined result are discussed in Overdetermined problems.
8.5 Controlling precision
Note: We consider the modification of solver precision to be an advanced feature, to be used sparingly, if
at all—and only after you have become comfortable building models in CVX.
Numerical methods for convex optimization are not exact; they compute their results to within a predefined
numerical precision or tolerance. Upon solution of your model, the tolerance level the solver has achieved
is returned in the cvx_slvtol variable. Attempts to interpret this tolerance level in any absolute sense
are not recommended. For one thing, each solver computes it differently. For another, it depends heavily on
the considerable transformations that CVX applies to your model before delivering it to the solver. So while
you may find its value interesting we strongly discourage dependence upon it within your applications.
The tolerance levels that CVX selects by default have been inherited from some of the underlying solvers
being used, with minor modifications. CVX actually considers three different tolerance levels solver ≤
standard ≤ reduced when solving a model:
• The solver tolerance solver is the level requested of the solver. The solver will stop as soon as it
achieves this level, or until no further progress is possible.
• The standard tolerance standard is the level at which CVX considers the model solved to full precision.
• The reduced tolerance reduced is the level at which CVX considers the model “inaccurately” sovled,
returning a status with the Inaccurate/ prefix. If this tolerance cannot be achieved, CVX returns
a status of Failed, and the values of the variables should not be considered reliable.
(See Interpreting the results for more information about the status messages.) Typically, solver = standard ,
but setting standard < solver has a useful interpretation: it allows the solver to search for more accurate
solutions without causing an Inaccurate/ or Failed condition if it cannot do so. The default values of
[solver , standard , reduced ] are set to [1/2 , 1/2 , 1/4 ], where = 2.22 × 10−16 is the machine precision. This
should be quite sufficient for most applications.
If you wish to modify the tolerances, you may do so using the cvx_precision command. There are
three ways to invoke this command. Called with no arguments, it will print the current tolerance levels to
the screen; or if called as a function, it will return those levels in a 3-element row vector.
Calling cvx_precision with a string argument allows you to select from a set of predefined precision
modes:
• cvx_precision low: [3/8 , 1/4 , 1/4 ]
• cvx_precision medium: [1/2 , 3/8 , 1/4 ]
• cvx_precision default: [1/2 , 1/2 , 1/4 ]
• cvx_precision high: [3/4 , 3/4 , 3/8 ]
• cvx_precision best: [0, 1/2 , 1/4 ]
48
Chapter 8. Solvers
The CVX Users’ Guide, Release 2.1
In function mode, these calls look like cvx_precision(’low’), etc. Note that the best precision
settings sets the solver target to zero, which means that the solver continues as long as it can make progress.
It will often be slower than default, but it is just as reliable, and sometimes produces more accurate
solutions.
Finally, the cvx_precision command can be called with a scalar, a length-2 vector, or a length-3 vector.
If you pass it a scalar, it will set the solver and standard tolerances to that value, and it will compute a
default reduced precision value for you. Roughly speaking, that reduced precision will be the square root
of the standard precision, with some bounds imposed to make sure that it stays reasonable. If you supply
two values, the smaller will be used for the solver and standard tolerances, and the larger for the reduced
tolerance. If you supply three values, their values will be sorted, and each tolerance will be set separately.
The cvx_precision command can be used either within a CVX model or outside of it; and its behavior
differs in each case. If you call it from within a model, e.g.,
cvx_begin
cvx_precision high
...
cvx_end
then the setting you choose will apply only until cvx_end is reached. If you call it outside a model, e.g.,
cvx_precision high
cvx_begin
...
cvx_end
then the setting you choose will apply globally; that is, to any subsequent models that are created and solved.
The local approach should be preferred in an application where multiple models are constructed and solved
at different levels of precision.
If you call cvx_precision in function mode, either with a string or a numeric value, it will return as its
output the previous precision vector—the same result you would obtain if you called it with no arguments.
This may seem confusing at first, but this is done so that you can save the previous value in a variable, and
restore it at the end of your calculations; e.g.,
cvxp = cvx_precision( 'high' );
cvx_begin
...
cvx_end
cvx_precision( cvxp );
This is considered good coding etiquette in a larger application where multiple CVX models at multiple precision levels may be employed. Of course, a simpler but equally courteous approach is to call
cvx_precision within the CVX model, as described above, so that its effect lasts only for that model.
8.5. Controlling precision
49
The CVX Users’ Guide, Release 2.1
8.6 Advanced solver settings
Warning: This is an advanced topic for users who have a deep understanding of the underlying solver
they are using, or who have received specific advice from the solver’s developer for improving performance. Improper use of the cvx_solver_settings command can cause unpredictable results.
Solvers can be tuned and adjusted in a variety of ways. Solver vendors attempt to select default settings that
will provide good performance across a broad range of problems. But no solver, and no choice of settings,
will perform well for every possible model. On occasion, it may be worthwhile to give a particular special
instructions to improve its performance for a specific application. Unfortunately, such settings differ from
solver to solver, so there is no way for CVX to provide this ability in a verifiable, reliable, global fashion.
Nevertheless, using the new cvx_solver_settings command, you can customize a solver’s settings
when a specific model demands it. We cannot emphasize enough that this is an expert feature to be employed
by experienced modelers only. Indeed, if you are an expert, you understand that these warnings are essential:
• CVX does not check the correctness of the settings you supply. If the solver rejects the settings, CVX
will fail until you change or remove those settings.
• There is no guarantee that altering the settings will improve performance in any way; indeed, it can
make the performance worse.
• CVX Research provides no documentation on the specific settings available for each solver; you will
have to consult the solver’s own documentation for this.
• The settings set here override any default values CVX may have chosen for each solver. Thus in
certain cases, using this feature this may actually confuse CVX and cause it to misinterpret the results.
For this reason, we cannot support all possible combinations of custom settings.
• Unless you have turned off solver output completely, CVX will warn you if any custom settings are
in effect every time you solve model.
With this warning out of the way, let us introduce cvx_solver_settings. Typing
cvx_solver_settings
at the command prompt provides a listing of the custom settings that have been provided for the active
solver. Custom settings are specific to each solver. Typing
cvx_solver_settings -all
will provide a full list of the custom settings provided for all solvers.
To create a new custom setting for the current solver, use this syntax:
cvx_solver_settings( '{name}', {value} )
{name} must be a valid MATLAB variable/field name. {value} can be any valid Matlab object; CVX
does not check its value in any way.
To clear all custom settings for the active solver, type
50
Chapter 8. Solvers
The CVX Users’ Guide, Release 2.1
cvx_solver_settings -clear
To clear just a single setting, type
cvx_solver_settings -clear {}
To clear all settings for all solvers, type
cvx_solver_settings -clearall
The settings created by the cvx_solver_settings command enjoy the same scope as cvx_solver,
cvx_precision, and so forth. For instance, if you use this command within a model—between
cvx_begin and cvx_end—the changes will apply only to that particular model. If you issue the command outside of a particular model, the change will persist through the end of the MATLAB session (or
until you change it again). Finally, if you use the cvx_save_prefs command, any custom settings you
have added will be saved and restored the next time you start Matlab.
8.6. Advanced solver settings
51
The CVX Users’ Guide, Release 2.1
52
Chapter 8. Solvers
CHAPTER
NINE
REFERENCE GUIDE
In this section we describe each operator, function, set, and command that you are likely to encounter in
CVX. In some cases, limitations of the underlying solver place certain restrictions or caveats on their use:
• Functions marked with a dagger (†) are not supported natively by the solvers that CVX uses. They
are handled using a successive approximation method which makes multiple calls to the underlying
solver, achieving the same final precision. If you use one of these functions, you will be warned
that successive approximation will be used. This technique is discussed further in The successive
approximation method. As this section discusses, this is an experimental approach that works well in
many cases, but cannot be guaranteed.
• Functions involving powers (e.g., x^p) and p-norms (e.g., norm(x,p)) are marked with a double
dagger (‡). CVX represents these functions exactly when p is a rational number. For irrational values
of p, a nearby rational is selected instead. See Power functions and p-norms for details on how both
cases are handled.
9.1 Arithmetic operators
Matlab’s standard arithmetic operations for addition +, subtraction -, multiplication, * .*, division / ./
\ .\, and exponentiation ^ .^ have been overloaded to work in CVX whenever appropriate—that is,
whenever their use is consistent with both standard mathematical and Matlab conventions and the DCP
ruleset. For example:
• Two CVX expressions can be added together if they are of the same dimension (or one is scalar) and
have the same curvature (i.e., both are convex, concave, or affine).
• A CVX expression can be multiplied or divided by a scalar constant. If the constant is positive, the
curvature is preserved; if it is negative, curvature is reversed.
• An affine column vector CVX expression can be multiplied by a constant matrix of appropriate dimensions; or it can be left-divided by a non-singular constant matrix of appropriate dimension.
Numerous other combinations are possible, of course. The use of the exponentiation operators ^ .^ are
somewhat limited; see the definitions of power in Nonlinear below.
Matlab’s basic matrix manipulation and arithmetic operations have been extended to work with CVX expressions as well, including:
• Concatenation: [ A, B ; C, D ]
53
The CVX Users’ Guide, Release 2.1
• Indexing: x(n+1:end), X([3,4],:), etc.
• Indexed assignment, including deletion: y(2:4) = 1, Z(1:4,:)
= [], etc.
• Transpose and conjugate transpose: Z.’, y’
9.2 Built-in functions
9.2.1 Linear
A number of Matlab’s basic linear and bilinear functions either work automatically with cvx expressions
or have been extended to do so, including: conj, conv, cumsum, diag, dot, find, fliplr, flipud,
flipdim, horzcat, hankel, ipermute, kron, mean, permute, repmat, reshape, rot90,
sparse, sum, trace, tril, triu, toeplitz, vertcat.
Most should behave identically with CVX expressions as they do with numeric expressions. Those that
perform some sort of summation, such as cumsum, sum, or multiplication, such as conv, dot or kron,
can only be used in accordance with the disciplined convex programming rules. For example, kron(X,Y)
is valid only if either X or Y is constant; and trace(Z) is valid only if the elements along the diagonal
have the same curvature.
9.2.2 Nonlinear
abs absolute value for real and complex arrays. Convex.
† exp exponential. Convex and nondecreasing.
† log logarithm. Concave and nondecreasing.
max maximum. Convex and nondecreasing.
min minimum. Concave and nondecreasing.
norm norms for real and complex vectors and matrices. Convex. Thus function follows the Matlab conventions closely. Thus the one-argument version norm(x) computes the 2-norm for vectors, and the
2-norm (maximum singular value) for matrices. The two-argument version norm(x,p) is supported
as follows:
• ‡ For vectors, all values p ≥ 1 are accepted.
• For matrices, p must be 1, 2, Inf, or ’Fro’.
polyval polynomial evaluation. polyval(p,x), where p is a vector of length n, computes
p(1) * x.^(n-1) + p(2) * x.^(n-2) + ... + p(n-1) * x + p(n)
This function can be used in CVX in two ways:
• If p is a variable and x is a constant, then polyval(x,p) computes a linear combination of
the elements of p. The combination must satisfy the DCP rules for addition and scaling.
54
Chapter 9. Reference guide
The CVX Users’ Guide, Release 2.1
• If p is a constant and x is a variable, then polyval(x,p) constructs a polynomial function
of the variable x. The polynomial must be affine, convex, or concave, and x must be real and
affine.
‡ power(x,p) x^p and x.^p, where x is a real variable and and p is a real constant. For x^p, both x
and p must be scalars. Only those values of p which can reasonably and unambiguously interpreted
as convex or concave are accepted:
• p = 0. Constant. x.^p is treated as identically 1.
• 0 < p < 1. Concave. The argument x must be concave (or affine), and is implicitly constrained
to be nonnegative.
• p = 1. Affine. x.^p is simply x.
• p ∈ {2, 4, 6, 8, ...}. Convex. Argument x must be affine.
• p > 1, p 6∈ {2, 3, 4, 5, ...}. Convex. Argument x must be affine, and is implicitly constrained to
be nonnegative.
Negative and odd integral values of p are not permitted, but see the functions pow_p, pow_pos, and
pow_abs in the next section for useful alternatives.
† power(p,x) p.^x and p^x, where p is a real constant and x is a real variable. For p^x, both p and x
must be scalars. Valid values of p include:
• p ∈ {0, 1}. Constant.
• 0 < p < 1. Convex and nonincreasing; x must be concave.
• p > 1. Convex and nondecreasing; x must be convex.
Negative values of p are not permitted.
std Standard deviation. Convex.
sqrt Square root. Implicitly constrains its argument to be nonnegative. Concave and nondecreasing.
var Variance. Convex.
9.3 New functions
Even though these functions were developed specifically for CVX, they work outside of a CVX specification
as well, when supplied with numeric arguments.
avg_abs_dev
The average absolute deviation about the mean µ(x) of x. Convex.
n
n
1X
1X
faad (x) =
|xi − µ(x)| =
xi −
n
n
i=1
i=1
1
n
P
i xi
=
1
(I − n1 11T )x
n
1
.
avg_abs_dev_med
9.3. New functions
55
The CVX Users’ Guide, Release 2.1
The average absolute deviation about the median m(x) of x. Convex.
n
n
i=1
i=1
1X
1X
|xi − m(x)| = inf
|xi − y|
faadm (x) =
y n
n
berhu(x,M)
The reversed Huber function (hence, Berhu), defined as
(
|x|
fberhu (x, M ) ,
(|x|2 + M 2 )/2M
|x| ≤ M
|x| ≥ M
Convex. If M is omitted, M = 1 is assumed; but if supplied, it must be a positive constant. Also
callable with three arguments as berhu(x,M,t), which computes t+t*berhu(x/t,M), useful
for concomitant scale estimation (see [Owen06]).
det_inv determinant of inverse of a symmetric (or Hermitian) positive definite matrix, det X −1 , which
is the same as the product of the inverses of the eigenvalues. When used inside a CVX specification,
det_inv constrains the matrix to be symmetric (if real) or Hermitian (if complex) and positive
semidefinite. When used with numerical arguments, det_inv returns +Inf if these constraints are
not met. Convex.
det_rootn n-th root of the determinant of a semidefinite matrix, (det X)1/n . When used inside a CVX
specification, det_rootn constrains the matrix to be symmetric (if real) or Hermitian (if complex)
and positive semidefinite. When used with numerical arguments, det_rootn returns -Inf if these
constraints are not met. Concave.
det_root2n the 2n-th root of the determinant of a semidefinite matrix;
i.e.,
det_root2n(X)=sqrt(det_rootn(X)). Concave. Maintained solely for back-compatibility
purposes.
† entr the elementwise entropy function: entr(x)=-x.*log(x). Concave. Returns -Inf when
called with a constant argument that has a negative entry.
Q
geo_mean the geometric mean of a vector, ( nk=1 xk )1/n . When used inside a CVX specification,
geo_mean constrains the elements of the vector to be nonnegative. When used with numerical
arguments, geo_mean returns -Inf if any element is negative. Concave and increasing.
huber(x,M)
The Huber function, defined as
(
|x|2
fhuber (x, M ) ,
2M |x| − M 2
|x| ≤ M
|x| ≥ M
Convex. If $x$ is a vector or array, the function is applied on an elementwise basis. If $M$ is
omitted, then $M=1$ is assumed; but if it supplied, it must be a positive constant. Also callable as
huber(x,M,t), which computes t+t*huber(x/t,M), useful for concomitant scale estimation
(see [Owen06]).
huber_circ(x,M)
56
Chapter 9. Reference guide
The CVX Users’ Guide, Release 2.1
The circularly symmetric Huber function, defined as
(
kxk22
fhuber_circ (x, M ) ,
2M kxk2 − M 2
kxk2 ≤ M
kxk2 ≥ M
Convex. Same (and implemented) as huber_pos(norm(x),M).
huber_pos(x,M) The same as the Huber function for nonnegative x; zero for negative x. Convex and
nondecreasing.
inv_pos The inverse of the positive portion, 1/ max{x, 0}. Inside CVX specification, imposes constraint
that its argument is positive. Outside CVX specification, returns +∞ if x ≤ 0. Convex and decreasing.
† kl_div Kullback-Leibler distance:
x log(x/y) − x + y
fkl (x, y) , 0
+∞
x, y > 0
x=y=0
otherwise
Convex. Outside CVX specification, returns +∞ if arguments aren’t in the domain.
lambda_max maximum eigenvalue of a real symmetric or complex Hermitian matrix. Inside CVX, imposes constraint that its argument is symmetric (if real) or Hermitian (if complex). Convex.
lambda_min minimum eigenvalue of a real symmetric or complex Hermitian matrix. Inside CVX, imposes constraint that its argument is symmetric (if real) or Hermitian (if complex). Concave.
lambda_sum_largest(X,k) sum of the largest k values of a real symmetric or complex Hermitian
matrix. Inside CVX, imposes constraint that its argument is symmetric (if real) or Hermitian (if
complex). Convex.
lambda_sum_smallest(X,k) sum of the smallest k values of a real symmetric or complex Hermitian
matrix. Inside CVX, imposes constraint that its argument is symmetric (if real) or Hermitian (if
complex). Concave.
log_det log of determinant of a positive definite matrix, log det(X). When used inside a CVX specification, log_det constrains its argument to be symmetric (if real) or Hermitian (if complex) and
positive definite. With numerical argument, log_det returns -Inf if these constraints are not met.
Concave.
‡ log_normcdf(x) logarithm of cumulative distribution function of standard normal random variable.
Concave and increasing. The current implementation is a fairly crude SDP-representable approximation, with modest accuracy over the interval [−4, 4]; we intend to replace it with a much better
approximation at some point.
Q
† log_prod(x) log i xi if when x is positive; −∞ otherwise. Concave and nonincreasing. Equivalent
to sum_log(x).
† log_sum_exp(x) the logarithm of the sum of the elementwise exponentials of x. Convex and nondecreasing.
logsumexp_sdp a polynomial approximation to the log-sum-exp function with global absolute accuracy.
This can be used to estimate the log-sum-exp function without using the successive approximation
method.
9.3. New functions
57
The CVX Users’ Guide, Release 2.1
matrix_frac(x,Y) matrix fractional function, xT Y −1 x. In CVX, imposes constraint that Y is symmetric (or Hermitian) and positive definite; outside CVX, returns +∞ unless Y = Y T 0. Convex.
norm_largest(x,k) For real and complex vectors, returns the sum of the largest k magnitudes in the
vector x. Convex.
norm_nuc(X) The sum of the singular values of a real or complex matrix X. (This is the dual of the usual
spectral matrix norm, i.e., the largest singular value.) Convex.
‡ norms(x,p,dim), norms_largest(x,k,dim) Computes vector norms along a specified dimension of a matrix or N-d array. Useful for sum-of-norms and max-of-norms problems. Convex.
poly_env(p,x) Computes the value of the convex or concave envelope of the polynomial described by
p (in the polyval sense). p must be a real constant vector whose length n is 0, 1, 2, 3, or some
other odd length; and x must be real and affine. The sign of the first nonzero element of p determines
whether a convex (positive) or concave (negative) envelope is constructed. For example, consider the
function p(x) , (x2 − 1)2 = x4 − 2x2 + 1, depicted along with its convex envelope in the figure
below.
The two coincide when |x| ≥ 1, but deviate when |x| < 1.
Attempting to
call polyval([1,0,2,0,1],x) in a CVX model would yield an error, but a call to
poly_env([1,0,2,0,1],x) yields a valid representation of the envelope. For convex or concave polynomials, this function produces the same result as polyval.
(x2−1)2
envelope
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
−0.2
−1.5
−1
−0.5
0
0.5
1
1.5
Figure 9.1: The polynomial function p(x) = x4 − 2x2 + 1 and its convex envelope.
58
Chapter 9. Reference guide
The CVX Users’ Guide, Release 2.1
pos(x) max{x, 0}, for real x. Convex and increasing.
‡ pow_abs(x,p) |x|p for x ∈ R or x ∈ C and p ≥ 1. Convex.
‡ pow_pos(x,p) max{x, 0}p for x ∈ R and p ≥ 1. Convex and nondecreasing.
‡ pow_p(x,p) for x ∈ R and real constant p, computes nonnegative convex and concave branches of the
power function:
(
xp
x>0
p≤0
fp (x) ,
convex, nonincreasing
(+∞ x ≤ 0
0 < p ≤ 1 fp (x) ,
p≥1
prod_inv(x)
Q
−1
i xi
fp (x) ,
xp
x≥0
(−∞ x < 0
xp
x≥0
+∞ x < 0
concave, nondecreasing
convex, nonmonotonic
when x is positive; +∞ otherwise. Convex and nonincreasing.
quad_form(x,P) xT P x for real x and symmetric P , and xH P x for complex x and Hermitian P . Convex in x for P constant and positive semidefinite; concave in x for P constant and negative semidefinite.
Note: Quadratic functions such as quad_form, sum_square can often be replaced by the norm function without sacrificing equivalence. For numerical reasons, this alternate formulation is preferred. Please
see Eliminating quadratic forms for more information.
quad_over_lin(x,y) xT x/y for x ∈ Rn , y > 0; for x ∈ Cn , y > 0, x∗ x/y. In CVX specification,
adds constraint that y > 0. Outside CVX specification, returns +∞ if y ≤ 0. Convex, and decreasing
in y.
quad_pos_over_lin(x,y) sum_square_pos( x )/y for x ∈ Rn , y > 0. Convex, increasing
in x, and decreasing in y.
† rel_entr(x) Scalar relative entropy; rel_entr(x,y)=x.*log(x/y). Convex.
sigma_max maximum singular value of real or complex matrix. Same as norm. Convex.
square x2 for x ∈ R. Convex.
square_abs |x|2 for x ∈ R or x ∈ C.
square_pos max{x, 0}2 for x ∈ R. Convex and increasing.
sum_largest(x,k) sum of the largest k values, for real vector x. Convex and nondecreasing.
P
† sum_log(x)
i log(xi ) when x is positive; −∞ otherwise. Concave and nondecreasing.
sum_smallest(x,k) sum of the smallest k values, i.e., equivalent to -sum_largest(-x,k). Concave and nondecreasing.
sum_square Equivalent to sum(square(x)), but more efficient. Convex. Works only for real values.
sum_square_abs Equivalent to sum(square_abs(x)), but more efficient. Convex.
9.3. New functions
59
The CVX Users’ Guide, Release 2.1
sum_square_pos Equivalent to sum(square_pos(x)), but more efficient. Works only for real values. Convex and increasing.
trace_inv(X) trace of the inverse of an SPD matrix X, which is the same as the sum of the inverses of
the eigenvalues. Convex. Outside of CVX, returns +Inf if argument is not positive definite.
trace_sqrtm(X) trace of the matrix square root of a positive semidefinite matrix X. which is the same
as the sum of the squareroots of the eigenvalues. Concave. Outside of CVX, returns +Inf if argument
is not positive semidefinite.
9.4 Sets
CVX currently supports the following sets; in each case, n is a positive integer constant.
nonnegative(n)
n
R+
, { x ∈ Rn | xi ≥ 0, i = 1, 2, . . . , n }
simplex(n)
n
R1+
, { x ∈ Rn | xi ≥ 0, i = 1, 2, . . . , n,
P
i xi
= 1}
lorentz(n)
Qn , { (x, y) ∈ Rn × R | kxk2 ≤ y }
rotated_lorentz(n)
Qnr , { (x, y, z) ∈ Rn × R × R | kxk2 ≤
√
yz, y, z ≥ 0 }
complex_lorentz(n)
Qnc , { (x, y) ∈ Cn × R | kxk2 ≤ y }
rotated_complex_lorentz(n)
Qnrc , { (x, y, z) ∈ Cn × R × R | kxk2 ≤
√
yz, y, z ≥ 0 }
semidefinite(n)
Sn+ ,
X ∈ Rn×n | X = X T , X 0
hermitian_semidefinite(n)
Hn+ ,
60
Z ∈ Cn×n | Z = Z H , Z 0
Chapter 9. Reference guide
The CVX Users’ Guide, Release 2.1
nonneg_poly_coeffs(n) The cone of all coefficients of nonnegative polynomials of degree n; n must
be even:
(
)
n
X
n
n−i
P+,n , p ∈ R [n + 1] |
pi+1 x
≥ 0 ∀x ∈ R
i=0
convex_poly_coeffs(n) The cone of all coefficients of convex polynomials of degree n; n must be
even:
(
)
n−2
X
n
n−i−2
P+,n , p ∈ R [n + 1] |
(n − i)(n − i − 1)pi+1 x
≥ 0 ∀x ∈ R
i=0
exp_cone
E , cl
n
(x, y, z) ∈ R × R × R | y > 0, yex/y ≤ z
o
geo_mean_cone(n)
(
Gn , cl
n
Y
(x, y) ∈ R × R × R | x ≥ 0, ( xi )1/n ≥ y
n
n
)
n
i=1
9.5 Commands
cvx_begin Begins a new CVX model. If a model is already in progress, it will issue a warning and clear
it. See cvx_begin and cvx_end for a full description, including the modifying keywords that control
solver output, SDP mode, GDP mode, etc.
cvx_clear Clears any model being constructed. Useful when an error has been made and it is necessary
to start from the beginning. Whereas cvx_begin issues a warning if called with a model in progress,
cvx_clear is silent.
cvx_end Signals the end of a CVX model. In typical use, this instructs CVX to begin the solution process.
See cvx_begin and cvx_end.
cvx_expert Controls the issuance of warnings when models requiring the use of successive approximation are employed; see The successive approximation method more details.
cvx_power_warning Controls if and when CVX issues warnings during the construction of models
involving rational power functions (i.e., x^p, where x is a variable and p is a constant); see Power
functions and p-norms.
cvx_precision Controls solver precision; see Controlling precision.
cvx_quiet Enables or disables screen output during the solution process; see Controlling screen output.
Also see cvx_begin and cvx_end for the newer, preferred syntax cvx_begin quiet.
cvx_save_prefs Saves the current states for cvx_expert, cvx_power_warning,
cvx_precision, and cvx_solver to disk, so that their values are retained when quitting
and re-starting MATLAB. The file is saved in MATLAB’s preference directory, which can be located
by typing the prefdir command.
9.5. Commands
61
The CVX Users’ Guide, Release 2.1
cvx_setup The setup script used to install and configure CVX; see Installation.
cvx_solver Selects the solver to be employed when solving CVX models; see Selecting a solver.
cvx_solver_settings Allows the user to deliver advanced, solver-specific settings to the solver that
CVX does not otherwise support; see Advanced solver settings.
cvx_version Prints information about the current versions of CVX, Matlab, and the operating system.
When submitting bug reports, please include the output of this command.
cvx_where Returns the directory where CVX is installed.
dual variable, dual variables Creates one or more dual variables to be connected to constraints
in the current model; see Dual variables.
expression, expressions Creates one or more expression holders; see Assignment and expression
holders.
maximise, maximize Specifies a maximization objective; see Objective functions.
minimise, minimize Specifies a minimization objective; see Objective functions.
variable, variables Creates one or more variables for use in the current CVX model; see Variables.
62
Chapter 9. Reference guide
CHAPTER
TEN
SUPPORT
The user base for CVX has grown to such an extent that full email-based support is no longer feasible for
our free version of CVX. Therefore, we have created several avenues for obtaining support.
For help on how to use CVX, this users’ guide is your first line of support. Please make sure that you have
attempted to find an answer to your question here before you pursue other avenues. We have placed this
document online and made it searchable in order to help you find the answers to the questions you may have.
With a package like CVX that encapsulates such mathematical complexity, it can sometimes be unclear if a
problem with a model is due to model formulation or a bug in CVX. See What is a bug? below to help you
discern the difference, and to determine the most appropriate channel for support.
10.1 The CVX Forum
If your answers cannot be found here, consider posting your question to the CVX Forum. This is a community forum that allows our users to submit questions and to answer other people’s questions. The forum
uses the open-source Askbot system, and its format should be familiar to anyone who participates in ORExchange, Stack Overflow, or any one of the Stack Exchange family of sites.
We highly encourage our expert users who enjoy helping others to participate in this forum. We hope that
it will not only serve as a resource for diagnosing problems and issues, but a clearinghouse for advanced
usage tips and tricks.
10.2 Bug reports
If you believe you have found a bug in CVX or in one of the underlying solvers, then we encourage you
to submit a bug report—either by email to to cvx@cvxr.com or through our web-based support portal.
Please include the following in your bug report so we can fully reproduce the problem:
1. the CVX model and supporting data that caused the error.
2. a copy of any error messages that it produced
3. the CVX version number and build number
4. the version number of Matlab that you are running
5. the name and version of the operating system you are using
63
The CVX Users’ Guide, Release 2.1
The easiest way to supply items 3-5 is to type cvx_version at the command prompt and copy its output
into your email message.
Please note that we do not own all of Matlab’s toolboxes. We cannot debug a model that employs functions
from a toolbox we do not own.
10.3 What is a bug?
Certain issues are unambiguously bugs, and you should feel free to report them immediately. In particular,
CVX often attempts to catch unexpected errors in key places—including cvx_setup, cvx_end, etc. It
will report those errors and instruct you to report them to us. If your model produces a MATLAB error that
CVX did not itself generate, and you cannot readily tie it to a syntax error in your model, please report that
as well.
That said, because disciplined convex programming is a new concept for many, we often receive support
requests for problems with their models that are not, in fact, bugs. A particularly common class of support
requests are reports of models being rejected due to Disciplined convex programming error
messages. For instance, the following code
variable x(10)
norm(x) == 1
will produce this error in CVX:
Error using cvxprob/newcnstr (line 181)
Disciplined convex programming error:
Invalid constraint: {convex} == {real constant}
Disciplined convex programming errors indicate that the model fails to adhere to the rules in the DCP
ruleset. In nearly all cases, this is not a bug, and should not be reported as such. Rather, the underlying issue
falls into one of two categories:
1. The model is not convex (mixed-integer or otherwise). A model with the nonlinear equation above
would fall squarely in this category. CVX simply cannot solve such problems. In some cases, it is
possible to transform a problem that is non-convex into one that is convex (for example, geometric
programs). This has not been the case for any problem submitted by our users—so far.
2. The model is convex, but it is still written in a manner that violates the rules. For instance, given the
same vector x above, the constraint
sqrt( sum( square( x ) ) ) <= 1
is convex, but it violates the ruleset—so it is rejected. However, the mathematically equivalent form
norm( x ) <= 1
is acceptable. If your error is of this type, you will need to find a way to express your problem in
a DCP-compliant manner. We have attempted to supply all of the commonly used functions that
the underlying solvers can support; so if you cannot easily rewrite your problem using the functions
supplied, it may not be possible. If you think this is a possibility, you may wish to see if the wizards
on the CVX Forum have suggestions for you.
64
Chapter 10. Support
The CVX Users’ Guide, Release 2.1
In rare cases, users have discovered that certain models were rejected with Disciplined convex
programming error messages even though they satisfied the DCP ruleset. We have not received a
bug report of this type in quite a long time, however, so we suspect our users have helped us iron out these
issues.
10.4 Handling numerical issues
No developer likes to tell their customers that their software may not work for them. Alas, we have no
choice. The fact is that we cannot guarantee that CVX will be able to solve your problem, even if it is
formulated properly, even if it avoids the use of integer or binary variables, even if it is of reasonable size,
even if it avoids the use of our experimental exponential cone support.
We blame the solvers—but we must come to their defense, too. Even the best and most mature solvers will
struggle with a particular problem that seems straightforward. Another solver may have no difficulty with
that one, but fail to find an accurate solution on another. While sometimes these challenges are due to bugs
in the solver’s implementation, quite often it is simply due to limits imposed by the nature of finite numerical
precision computation. And different problems push those limits to different degrees. So the fact is that no
solver is perfect, but no solver can be.
When we consider mixed-integer problems, the situation is even worse. Solvers must perform what is
effectively an exhaustive search among the integer variables to determine the correct solution. Yes, there are
some intelligent and innovative ways to speed up that search, and the performance of mixed-integer solvers
has improved dramatically over the years. But there will always be models for which the exhaustive search
will simply take too long.
None of this is much comfort if it is your model the solver is struggling with. Here are some practical tips if
you encounter this problem:
Try a different solver. Use the cvx_solver command for this. If you are using Gurobi or MOSEK, don’t
hesitate to try one of the free solvers if they are compatible with your problem.
Reduce the precision. Consider inserting cvx_precision medium or even cvx_precision low
into your problem to see if that allows the solver to exit successfully. Of course, if it does succeed,
make sure to check the results to see if they are acceptable for your application. If they are not,
consider some of the other advice here to see if the solvability of your model may be improved.
Remove constraints. If you think that one or more of the constraints might not be active at the solution, try
removing them. If the solver terminates, you can confirm that your guess was correct by examining
the solution to the modified problem.
Add constraints. Consider adding simple bounds to the constraints to reduce the size of the feasible set.
This will sometimes improve the numerical conditioning of the problem. Make them as tight as you
can without impinging on the optimal set. If this modified problem is successfully solved, check the
solution to see if any of the added bounds are active. If they are, relax them, and try again.
Watch for scaling issues. Scaling issues are the most vexing problems for numerical solvers to deal with.
Solvers will often re-scale the problem to reduce the dynamic range of the numerical coefficients, but
doing so sometimes leads to undesirable effects on the solution. It is better to avoid scaling issues
during the modeling process. Don’t mix values of wildly different magnitudes, such as 1e-3 and
10.4. Handling numerical issues
65
The CVX Users’ Guide, Release 2.1
1e20. Even better, try to avoid any numerical values (both in fixed parameters and likely values of
the variables) that exceed 1e8 in absolute value.
Try equivalent reformulations. It is quite likely that your model can be expressed in a variety of different
ways. Certainly, you should begin with the most obvious and natural formulation; but if you encounter
numerical issues, a reformulation may often solve them. One reformulation we highly recommend is
to eliminate quadratic forms; see this section for more details.
Reach out to the CVX Forum. Share your struggles with the larger CVX community! Perhaps they will
have concrete suggestions for improving the solvability of your model.
10.5 CVX Professional support
Paid CVX Professional users will receive support through a trouble-ticket support system, so that they can
have confidence that their issues are being addressed promptly. This infrastructure is still under development; we will update this section and the Web site with more information once it has been completed.
66
Chapter 10. Support
CHAPTER
ELEVEN
ADVANCED TOPICS
Note: In this section we describe a number of the more advanced capabilities of CVX. We recommend that
you skip this section at first, until you are comfortable with the basic capabilities described above.
11.1 Eliminating quadratic forms
One particular reformulation that we strongly encourage is to eliminate quadratic forms—that is, functions
like sum_square, sum(square(.)) or quad_form—whenever it is possible to construct equivalent
models using norm instead. Our experience tells us that quadratic forms often pose a numerical challenge
for the underlying solvers that CVX uses.
We acknowledge that this advice goes against conventional wisdom: quadratic forms are the prototypical
smooth convex function, while norms are nonsmooth and therefore unwieldy. But with the conic solvers
that CVX uses, this wisdom is exactly backwards. It is the norm that is best suited for conic formulation
and solution. Quadratic forms are handled by converting them to a conic form—using norms, in fact! This
conversion process poses some interesting scaling challenges. It is better if the modeler can eliminate the
need to perform this conversion.
For a simple example of such a change, consider the objective
minimize( sum_square( A * x - b ) )
In exact arithmetic, this is precisely equivalent to
minimize( square_pos( norm( A * x - b ) ) )
But equivalence is also preserved if we eliminate the square altogether:
minimize( norm( A * x - b ) )
The optimal value of x is identical in all three cases, but this last version is likely to produce more accurate
results. Of course, if you need the value of the squared norm, you can always recover it by squaring the
norm after the fact.
Conversions using quad_form can sometimes be a bit more difficult. For instance, consider
quad_form( A * x - b, Q ) <= 1
67
The CVX Users’ Guide, Release 2.1
where Q is a positive definite matrix. The equivalent norm version is
norm( Qsqrt * ( A * x - b ) ) <= 1
where Qsqrt is an appropriate matrix square root of Q. One option is to compute the symmetric square root
Qsqrt = sqrtm(Q), but this computation destroys sparsity. If Q is sparse, it is likely worth the effort to
compute a sparse Cholesky-based square root:
[ Qsqrt, p, S ] = chol( Q );
Qsqrt = Qsqrt * S;
Sometimes an effective reformulation requires a practical understanding of what it means for problems to
be equivalent. For instance, suppose we wanted to add an `1 regularization term to the objective above,
weighted by some fixed, positive lambda:
minimize( sum_square( A * x - b ) + lambda * norm( x, 1 ) )
In this case, we typically do not care about the specific values of lambda; rather we are varying it over a
range to study the tradeoff between the residual of A*x-b and the 1-norm of x. The same tradeoff can be
studied by examining this modified model:
minimize( norm( A * x - b ) + lambda2 * norm( x, 1 ) )
This is not precisely the same model; setting lambda and lambda2 to the same value will not yield
identical values of x. But both models do trace the same tradeoff curve—only the second form is likely to
produce more accurate results.
11.2 Indexed dual variables
In some models, the number of constraints depends on the model parameters—not just their sizes. It is
straightforward to build such models in CVX using, say, a Matlab for loop. In order to assign each of these
constraints a separate dual variable, we must find a way to adjust the number of dual variables as well. For
this reason, CVX supports indexed dual variables. In reality, they are simply standard Matlab cell arrays
whose entries are CVX dual variable objects.
Let us illustrate by example how to declare and use indexed dual variables. Consider the following semidefinite program from the SeDuMi examples:
P
minimize Pni=1 (n − i)Xii
n
subject to
i=1 Xi,i+k = bk , k = 1, 2, . . . , n
X0
This problem minimizes a weighted sum of the main diagonal of a positive semidefinite matrix, while
holding the sums along each diagonal constant. The parameters of the problem are the elements of the
vector b ∈ Rn , and the optimization variable is a symmetric matrix X ∈ Rn×n . The CVX version of this
model is
cvx_begin
variable X( n, n ) symmetric
minimize( ( n - 1 : -1 : 0 ) * diag( X ) );
for k = 0 : n-1,
68
Chapter 11. Advanced topics
The CVX Users’ Guide, Release 2.1
sum( diag( X, k ) ) == b( k+1 );
end
X == semidefinite(n);
cvx_end
If we wish to obtain dual information for the n simple equality constraints, we need a way to assign each
constraint in the for loop a separate dual variable. This is accomplished as follows:
cvx_begin
variable X( n, n ) symmetric
dual variables y{n}
minimize( ( n - 1 : -1 : 0 ) * diag( X ) );
for k = 0 : n-1,
sum( diag( X, k ) ) == b( k+1 ) : y{k+1};
end
X == semidefinite(n);
cvx_end
The statement dual variables y{n} allocates a cell array of n dual variables, and stores the result
in the Matlab variable Z. The equality constraint in the for loop has been augmented with a reference to
y{k+1}, so that each constraint is assigned a separate dual variable. When the cvx_end command is
issued, CVX will compute the optimal values of these dual variables, and deposit them into an n-element
cell array y.
This example admittedly is a bit simplistic. With a bit of careful arrangement, it is possible to rewrite this
model so that the n equality constraints can be combined into a single vector constraint, which in turn would
require only a single vector dual variable. 1 For a more complex example that is not amenable to such a
simplification, see the file
examples/cvxbook/Ch07_statistical_estim/cheb.m
in the CVX distribution. In that problem, each constraint in the for loop is a linear matrix inequality, not a
scalar linear equation; so the indexed dual variables are symmetric matrices, not scalars.
11.3 The successive approximation method
Note: If you were referred to this web page by CVX’s warning message: welcome! Please read this section
carefully to fully understand why using functions like log, exp, etc. within CVX models requires special
care.
Prior to version 1.2, the functions exp, log, log_det, and other functions from the exponential family
could not be used within CVX. Unfortunately, CVX utilizes symmetric primal/dual solvers that simply
cannot support those functions natively 2 , and a variety of factors prevent us from incorporating other types
1
Indeed, a future version of CVX will support the use of the Matlab function spdiags, which will reduce the entire for loop
to the single constraint spdiags(X,0:n-1)==b.
2
Technically there are a couple of exceptions here. First of all, SDPT3 does, in fact, support the existence of logarithms and
log_det terms in the objective function. However, it doesn’t support such terms within constraints. Unfortunately, because CVX
does not differentiate between objective terms and constraint terms internally, it is not able to utilize this capability of SDPT3.
Secondly, this section was written before the inclusion of MOSEK support in CVX, and CVX does indeed provide support for
11.3. The successive approximation method
69
The CVX Users’ Guide, Release 2.1
of solvers into CVX.
Nevertheless, support for these functions was requested quite frequently. For this reason, we constructed a
successive approximation heuristic that allows the symmetric primal/dual solvers to support the exponential
family of functions. A precise description of the approach is beyond the scope of this text, but roughly
speaking, the method proceeds as follows:
1. Choose an initial approximation centerpoint xc = 0.
2. Construct a polynomial approximation for each log/exp/entropy term which is accurate in the neighborhood of xc .
3. Solve this approximate model to obtain its optimal point x̄.
4. If x̄ satisfies the optimality conditions for the orignal model to sufficient precision, exit.
5. Otherwise, shift xc towards x̄, and repeat steps 2-5.
Again, this is a highly simplified description of the approach; for instance, we actually employ both the
primal and dual solutions to guide our judgements for shifting xc and terminating.
This approach has proven surprisingly effective for many problems. However, as with many heuristic approaches, it is not perfect. It will sometimes fail to converge even for problems known to have solutions.
Even when it does converge, it is several times slower than the standard solver, due to its iterative approach.
Therefore, it is best to use it sparingly and carefully. Here are some specific usage tips:
• First, confirm that the log/exp/entropy terms are truly necessary for your model. In many cases, an
exactly equivalent model can be constructed without them, and that should always be preferred. For
instance, the constraint
sum_log(x) >= 10
can be expressed in terms of the geo_mean function as
geo_mean(x) >= log(10)^(1/length(x))
Many determinant maximization problems are commonly written using log_det, but in fact that is
often unnecessary. For instance, consider the objective
minimize( log_det(X) )
CVX actually converts this internally to this:
minimize( n*log(det_rootn(X)) )
So what you can do instead is simply remove the logarithm, and solve this instead:
minimize( det_rootn(X) )
The value of log_det(X) can simply be computed after the model is completed. Unfortunately,
this only works if log_det is the only term in the objective; so, for instance, this function cannot,
unfortunately, be converted in a similar fashion:
smooth nonlinearities in its solver. But this capability is not easy to use in MATLAB.
70
Chapter 11. Advanced topics
The CVX Users’ Guide, Release 2.1
minimize( log_det(X) + trace(C*X) )
• Second, try different solvers. For instance, SeDuMi and MOSEK tend to be more effective with the
successive approximation method than SDPT3. So if the default solver choice fails to give a solution
to your model, try switching to one of these solvers.
• Third, try smaller instances of your problem. If they succeed where the larger instance fails, then at
least you can confirm if the model is behaving as you hope before considering alternative options like
a different solver.
The bottom line, unfortunately, is that we cannot guarantee that the successive approximation approach will
successfully handle your specific models. If you encounter problems, you are invited to submit a bug report,
but we will not be able to promise a fix.
11.3.1 Suppressing the warning
Because of all of these caveats, we believe that it is necessary to issue a warning when it is used so that
users understand its experimental nature. This warning appears the first time you attempt to specify a model
in CVX that uses an function that requires the successive approximation method. In fact, that warning may
very well have brought you to this section of the manual.
If you wish to suppress this warning in the future, simply issue the command
cvx_expert true
before you construct your model. If you wish to suppress this message for all future sessions of MATLAB,
follow this command with the cvx_save_prefs command.
11.4 Power functions and p-norms
In order to implement the convex or concave branches of the power function xp and p-norms kxkp for
general values of p, CVX uses an enhanced version of the SDP/SOCP conversion method described by
[AG00]. This approach is exact—as long as the exponent p is rational. To determine integral values pn , pd
such that pn /pd = p, CVX uses Matlab’s rat function with its default tolerance of 10−6 . There is currently
no way to change this tolerance. See the MATLAB documentation for the rat function for more details.
The complexity of the resulting model depends roughly on the size of the values pn and pd . Let us introduce
a more precise measure of this complexity. For p = 2, a constraint xp ≤ y can be represented with exactly
one 2 × 2 LMI:
y x
2
x ≤ y ⇐⇒
0.
x 1
For other values of p = pn /pd , CVX generates a number of 2 × 2 LMIs that depends on both pn and pd ;
we denote this number by k(pn , pd ). (In some cases additional linear constraints are also generated, but we
ignore them for this analysis.) For instance, for p = 3/1, we have
z x
y z
3
x ≤ y, x ≥ 0 ⇐⇒ ∃z
0.
0.
x 1
z x
11.4. Power functions and p-norms
71
The CVX Users’ Guide, Release 2.1
So k(3, 1) = 2. An empirical study has shown that for p = pn /pd > 1, we have
k(pn , pd ) ≤ log2 pn + α(pn )
where the α(pn ) term grows very slowly compared to the log2 term. Indeed, for pn ≤ 4096, we have verified
that α(pn ) is usually 1 or 2, but occasionally 0 or 3. Similar results are obtained for 0 < p < 1 and p < 0.
The cost of this SDP representation is relatively small for nearly all useful values of p. Nevertheless,
CVX issues a warning whenever k(pn , pd ) > 10 to insure that the user is not surprised by any unexpected slowdown. In the event that this threshold does not suit you, you may change it using the command cvx_power_warning(thresh), where thresh is the desired cutoff value. Setting the threshold to Inf disables it completely. As with the command cvx_precision, you can place a call to
cvx_power_warning within a model to change the threshold for a single model; or outside of a model
to make a global change. The command always returns the previous value of the threshold, so you can save
it and restore it upon completion of your model, if you wish. You can query the current value by calling
cvx_power_warning with no arguments.
11.5 Overdetermined problems
The status message Overdetermined commonly occurs when structure in a variable or set is not properly
recognized. For example, consider the problem of finding the smallest diagonal addition to a matrix W ∈
Rn×n to make it positive semidefinite:
minimize Tr(D)
subject to W + D 0
D diagonal
In CVX, this problem might be expressed as follows:
n = size(W,1);
cvx_begin
variable D(n,n) diagonal;
minimize( trace( D ) );
subject to
W + D == semidefinite(n);
cvx_end
If we apply this specification to the matrix W=randn(5,5), a warning is issued,
Warning: Overdetermined equality constraints;
problem is likely infeasible.
and the variable cvx_status is set to Overdetermined.
What has happened here is that the unnamed variable returned by statement semidefinite(n) is symmetric, but W is fixed and unsymmetric. Thus the problem, as stated, is infeasible. But there are also n2
equality constraints here, and only n + n ∗ (n + 1)/2 unique degrees of freedom—thus the problem is
overdetermined. We can correct this problem by replacing the equality constraint with
72
Chapter 11. Advanced topics
The CVX Users’ Guide, Release 2.1
sym( W ) + D == semidefinite(n);
sym is a function we have provided that extracts the symmetric part of its argument; that is, sym(W) equals
0.5 * ( W + W’ ).
11.6 Adding new functions to the atom library
CVX allows new convex and concave functions to be defined and added to the atom library, in two ways,
described in this section. The first method is simple, and can (and should) be used by many users of CVX,
since it requires only a knowledge of the basic DCP ruleset. The second method is very powerful, but a bit
complicated, and should be considered an advanced technique, to be attempted only by those who are truly
comfortable with convex analysis, disciplined convex programming, and CVX in its current state.
Please let us know if you have implemented a convex or concave function that you think would be useful to
other users; we will be happy to incorporate it in a future release.
11.6.1 New functions via the DCP ruleset
The simplest way to construct a new function that works within CVX is to construct it using expressions
that fully conform to the DCP ruleset. Consider, for instance, the deadzone function
(
0
|x| ≤ 1
f (x) = max{|x| − 1, 0} =
x−1 x>1
To implement this function in CVX, simply create a file deadzone.m containing
function y = deadzone( x )
y = max( abs( x ) - 1, 0 )
This function works just as you expect it would outside of CVX — that is, when its argument is numerical.
But thanks to Matlab’s operator overloading capability, it will also work within CVX if called with an affine
argument. CVX will properly conclude that the function is convex, because all of the operations carried out
conform to the rules of DCP: abs is recognized as a convex function; we can subtract a constant from it,
and we can take the maximum of the result and 0, which yields a convex function. So we are free to use
deadzone anywhere in a CVX specification that we might use abs, for example, because CVX knows
that it is a convex function.
Let us emphasize that when defining a function this way, the expressions you use must conform to the DCP
ruleset, just as they would if they had been inserted directly into a CVX model. For example, if we replace
max with min above; e.g.,
function y = deadzone_bad( x )
y = min( abs( x ) - 1, 0 )
then the modified function fails to satisfy the DCP ruleset. The function will work outside of a CVX
specification, happily computing the value min{|x| − 1, 0} for a numerical argument x. But inside a CVX
specification, invoked with a nonconstant argument, it will generate an error.
11.6. Adding new functions to the atom library
73
The CVX Users’ Guide, Release 2.1
11.6.2 New functions via partially specified problems
A more advanced method for defining new functions in CVX relies on the following basic result of convex
analysis. Suppose that S ⊂ Rn × Rm is a convex set and g : (Rn × Rm ) → (R ∪ +∞) is a convex
function. Then
f : Rn → (R ∪ +∞),
f (x) , inf { g(x, y) | ∃y, (x, y) ∈ S }
is also a convex function. (This rule is sometimes called the partial minimization rule.) We can think of the
convex function f as the optimal value of a family of convex optimization problems, indexed or parametrized
by x,
minimize g(x, y)
subject to (x, y) ∈ S
with optimization variable y.
One special case should be very familiar: if m = 1 and g(x, y) , y, then
f (x) , inf { y | ∃y, (x, y) ∈ S }
gives the classic epigraph representation of f :
epi f = S + ({0} × R+ ) ,
where 0 ∈ Rn .
In CVX you can define a convex function in this very manner, that is, as the optimal value of a parameterized
family of disciplined convex programs. We call the underlying convex program in such cases an incomplete
specification—so named because the parameters (that is, the function inputs) are unknown when the specification is constructed. The concept of incomplete specifications can at first seem a bit complicated, but it is
very powerful mechanism that allows CVX to support a wide variety of functions.
Let us look at an example to see how this works. Consider the unit-halfwidth Huber penalty function h(x):
(
x2
|x| ≤ 1
h : R → R, h(x) ,
.
2|x| − 1 |x| ≥ 1
We can express the Huber function in terms of the following family of convex QPs, parameterized by x:
minimize 2v + w2
subject to |x| ≤ v + w
w ≤ 1, v ≥ 0
with scalar variables v and w. The optimal value of this simple QP is equal to the Huber penalty function of
x. We note that the objective and constraint functions in this QP are (jointly) convex in v, w and x.
We can implement the Huber penalty function in CVX as follows:
function cvx_optval = huber( x )
cvx_begin
variables w v;
minimize( w^2 + 2 * v );
subject to
abs( x ) <= w + v;
w <= 1; v >= 0;
cvx_end
74
Chapter 11. Advanced topics
The CVX Users’ Guide, Release 2.1
If huber is called with a numeric value of x, then upon reaching the cvx_end statement, CVX will find
a complete specification, and solve the problem to compute the result. CVX places the optimal objective
function value into the variable cvx_optval, and function returns that value as its output. Of course, it’s
very inefficient to compute the Huber function of a numeric value x by solving a QP. But it does give the
correct value (up to the core solver accuracy).
What is most important, however, is that if huber is used within a CVX specification, with an affine CVX
expression for its argument, then CVX will do the right thing. In particular, CVX will recognize the Huber
function, called with affine argument, as a valid convex expression. In this case, the function huber will
contain a special Matlab object that represents the function call in constraints and objectives. Thus the
function huber can be used anywhere a traditional convex function can be used, in constraints or objective
functions, in accordance with the DCP ruleset.
There is a corresponding development for concave functions as well. Given a convex set S as above, and a
concave function g : (Rn × Rm ) → (R ∪ −∞), the function
f : R → (R ∪ −∞),
f (x) , sup { g(x, y) | ∃y, (x, y) ∈ S }
is concave. If g(x, y) , y, then
f (x) , sup { y | ∃y, (x, y) ∈ S }
gives the hypograph representation of f :
hypo f = S − Rn+ .
In CVX, a concave incomplete specification is simply one that uses a maximize objective instead of a
minimize objective; and if properly constructed, it can be used anywhere a traditional concave function
can be used within a CVX specification.
For an example of a concave incomplete specification, consider the function
f : Rn×n → R,
f (X) = λmin (X + X T )
Its hypograph can be represented using a single linear matrix inequality:
hypo f = { (X, t) | f (X) ≥ t } = (X, t) | X + X T − tI 0
So we can implement this function in CVX as follows:
function cvx_optval = lambda_min_symm( X )
n = size( X, 1 );
cvx_begin
variable y;
maximize( y );
subject to
X + X' - y * eye( n ) == semidefinite( n );
cvx_end
If a numeric value of X is supplied, this function will return min(eig(X+X’)) (to within numerical
tolerances). However, this function can also be used in CVX constraints and objectives, just like any other
concave function in the atom library.
11.6. Adding new functions to the atom library
75
The CVX Users’ Guide, Release 2.1
There are two practical issues that arise when defining functions using incomplete specifications, both of
which we will illustrate using our huber example above. First of all, as written the function works only
with scalar values. To apply it (elementwise) to a vector requires that we iterate through the elements in a
for loop—a very inefficient enterprise, particularly in CVX. A far better approach is to extend the huber
function to handle vector inputs. This is, in fact, rather simple to do: we simply create a multiobjective
version of the problem:
function cvx_optval = huber( x )
sx = size( x );
cvx_begin
variables w( sx ) v( sx );
minimize( w .^ 2 + 2 * v );
subject to
abs( x ) <= w + v;
w <= 1; v >= 0;
cvx_end
This version of huber will in effect create sx “instances” of the problem in parallel; and when used in a
CVX specification, will be handled correctly.
The second issue is that if the input to huber is numeric, then direct computation is a far more efficient
way to compute the result than solving a QP. (What is more, the multiobjective version cannot be used with
numeric inputs.) One solution is to place both versions in one file, with an appropriate test to select the
proper version to use:
function cvx_optval = huber( x )
if isnumeric( x ),
xa
= abs( x );
flag = xa < 1;
cvx_optval = flag .* xa.^2 + (~flag) * (2*xa-1);
else,
sx = size( x );
cvx_begin
variables w( sx ) v( sx );
minimize( w .^ 2 + 2 * v );
subject to
abs( x ) <= w + v;
w <= 1; v >= 0;
cvx_end
end
Alternatively, you can create two separate versions of the function, one for numeric input and one for CVX
expressions, and place the CVX version in a subdirectory called @cvx. (Do not include this directory in
your Matlab path; only include its parent.) Matlab will automatically call the version in the @cvx directory
when one of the arguments is a CVX variable. This is the approach taken for the version of huber found
in the CVX atom library.
One good way to learn more about using incomplete specifications is to examine some of the examples
already in the CVX atom library. Good choices include huber, inv_pos, lambda_min, lambda_max,
matrix_frac, quad_over_lin, sum_largest, and others. Some are a bit difficult to read because
of diagnostic or error-checking code, but these are relatively simple.
76
Chapter 11. Advanced topics
CHAPTER
TWELVE
LICENSE
CVX: A system for disciplined convex programming
© 2012 CVX Research, Inc., Austin, TX.
http://cvxr.com
info@cvxr.com
Thank you for using CVX!
The files contained in the CVX distribution come from several different sources and are covered under a
variety of licenses. Those files owned by CVX Research, Inc. are released under one of two licenses: the
CVX Standard License and the CVX Professional License. The CVX Professional License is not a permissive
license, and it does not permit redistribution. The CVX Standard License, on the other hand, is in fact the
GNU Standard License, Version 3, amended with a single additional permission.
12.1 CVX Professional License
The full CVX package, including the commercial solver functionality, includes several files in Matlab pcode format. This format contains an encrypted form of Matlab bytecode, and are distinguished by a .p
suffix. For instance, the package currently includes the files
shims/cvx_mosek.p
shims/cvx_gurobi.p
cvx_license.p
but future versions may add additional .p files, and they are governed by this text as well. Furthermore, all
files within the subdirectories
gurobi/
mosek/
are licensed by Gurobi Optimization and MOSEK ApS, respectively, for inclusion in the CVX package by
contractual agreement. In the future, CVX Research may include other commercial solvers in its distributions, and they will be governed by this section as well.
Any package containing even one of the files mentioned above is subject to the CVX Professional License.
You may not redistribute such a package, nor make it available on a public repository, without the express
written consent of CVX Research, Inc. Of course, the full CVX package is freely available to download
77
The CVX Users’ Guide, Release 2.1
directly from our servers, so please feel free to refer people to our web site—even directly to our download
page.
12.2 CVX Standard License
CVX also offers what we call “redistributable” packages that omit all of the files listed above. This package
retains all of the functionality of CVX, with the exception of its ability to connect to commercial solvers.
Redistributable packages are offered on the same download page as our standard packages.
The redistributable package is covered by the CVX Standard License. This license is nothing more than
the GNU Standard License, Version 3 (GPLv3), with one additional permission we call the “Free Solver
Clause”, discused in the next section.
The full text of the GPLv3 is given in the file GPL.txt found in the top directory of the CVX package. Please
see Section 7 of this license for more information about additional permissions such as the one offered here.
In particular, you are free to remove this permission from any modified version of CVX you create; but you
may not create new ones.
You must include the files LICENSE.txt and GPL.txt in unmodified form when redistributing this software
or any subset of it. If you did not receive a copy of either of these files with your distribution, please contact
us.
12.3 The Free Solver Clause
We have chosen to add a single “additional permission” clause to the CVX Standard License above and
beyond the text of the GPLv3. This full text of this clause is contained in the following paragraph:
You are free to modify CVX to connect to any solver whose full functionality is provided free of
charge to all users, even if the solver’s license is otherwise incompatible with the GNU Standard
License Version 3. This permission does not apply to solvers which are free just to certain users
(e.g., academic users), free for limited time periods, or free with restricted functionality.
CVX relies upon other software packages, called solvers, to perform many of its underlying calculations.
Currently CVX supports free solvers SeDuMi and SDPT3, and commercial solvers Gurobi and MOSEK.
The resulting nexus of free and commercial software presents a licensing challenge. Our vision is guided by
three goals:
• to ensure that CVX remains fully free to use with any fully free solver.
• to support the development and support of CVX by selling interfaces to commercial solvers to commercial customers.
• to provide the academic community with our full commercial capabilities at no charge. (Obviously
this depends upon the cooperation of commercial solver vendors. Gurobi and MOSEK fully support
this goal.)
The terms we lay out here are intended to support these trifold goals.
We invite our users to create new interfaces between CVX and other free solvers. By “free”, we mean
that the solver must be made available at no charge for to all users, including commercial users, without
78
Chapter 12. License
The CVX Users’ Guide, Release 2.1
restriction. Please contact us if you are interested in creating such an interface; we can offer assistance. If
you do create one, please consider submitting it to us for inclusion in the standard CVX distribution; but
you are under no obligation to do this. Instead, you can ship the interface code with the solver itself; or you
can construct a modified version of the redistributable CVX package with your interface included.
We do not permit the creation and distribution of new interfaces between CVX and non-free solvers—even
if those solvers are made available to academic academic users at no charge. If you are a vendor or developer
of a commercial solver, and would like to develop or offer a CVX interface to your users, please contact
us at info@cvxr.com. We welcome the opportunity to support a wider variety of commercial solvers with
CVX, and are willing to devote engineering resources to make those connections. Of course, we hope that
you support our goal of making commercial solver capability available at no charge to academic customers;
and our licensing interface helps to facilitate that offering.
If you are a user of a particular commercial solver and would like to see it supported by CVX, please
contact your solver vendor—but please contact us at info@cvxr.com as well. If there is sufficient demand,
and it proves technically and financially feasible, we will reach out to the solver vendor to work on an
implementation.
12.4 Bundled solvers
The solvers SDPT3, SeDuMi, Gurobi, and MOSEK are distributed with CVX in the sdpt3/, sedumi/,
gurobi/, and mosek/ subdirectories, respectively. (The redistributable version of CVX does not include
gurobi/ and mosek/.) Future versions of CVX may include additional solvers as well, each in their
own subdirectories. None of these packages are owned by CVX Research, Inc. They are included with
permission of the authors. The solvers SDPT3 and SeDuMi are subject to the terms of the GPLv2; therefore,
you are free to redistribute them under the terms of that license.
12.5 Example library
The contents of the example library, which is distributed with CVX in the examples/ subdirectory, is public
domain. You are free to use them in any way you wish; but when you do, we request that you give appropriate
credit to the authors. A number of people have contributed to the examples in this library, including Lieven
Vandenberghe, Joëlle Skaf, Argyris Zymnis, Almir Mutapcic, Michael Grant, and Stephen Boyd. If you
would like to offer some examples to be included in this library, and are willing to release them to the public
domain, please contact us!
12.6 No Warranty
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without
even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
12.4. Bundled solvers
79
The CVX Users’ Guide, Release 2.1
80
Chapter 12. License
CHAPTER
THIRTEEN
CITING CVX
If you are actively using CVX in teaching, research, or applications, and haven’t yet told us about it, please
do so! Drop us an email at CVX Research Support. It is truly encouraging to hear about new uses for CVX,
and we like to keep track of geographic and technical diversity of our user base. And of course, it is always
a pleasure to receive an email at CVX Research Support that is not a bug report!
Are you using CVX in research work to be published? If so, please include explicit mention of our work in
your publication. We suggest language such as this:
To solve problem (17) we used CVX, a package for specifying and solving convex programs
[1],[2].
with the following corresponding entries in your bibliography:
CVX Research, Inc. CVX: Matlab software for disciplined convex programming, version 2.0.
http://cvxr.com/cvx, April 2011.
M. Grant and S. Boyd. Graph implementations for nonsmooth convex programs, Recent Advances in Learning and Control (a tribute to M. Vidyasagar), V. Blondel, S. Boyd, and H.
Kimura, editors, pages 95-110, Lecture Notes in Control and Information Sciences, Springer,
2008. http://stanford.edu/~boyd/graph_dcp.html.
The corresponding BiBTeX citations are given below:
@misc{cvx,
author
= {CVX Research, Inc.},
title
= {{CVX}: Matlab Software for Disciplined Convex Programming, version 2.0},
howpublished = {\url{http://cvxr.com/cvx}},
month
= aug,
year
= 2012
}
@incollection{gb08,
author
= {M. Grant and S. Boyd},
title
= {Graph implementations for nonsmooth convex programs},
booktitle = {Recent Advances in Learning and Control},
series
= {Lecture Notes in Control and Information Sciences},
editor
= {V. Blondel and S. Boyd and H. Kimura},
publisher = {Springer-Verlag Limited},
pages
= {95--110},
year
= 2008,
note
= {\url{http://stanford.edu/~boyd/graph_dcp.html}}
}
81
The CVX Users’ Guide, Release 2.1
82
Chapter 13. Citing CVX
CHAPTER
FOURTEEN
CREDITS AND ACKNOWLEDGEMENTS
CVX was designed by Michael Grant and Stephen Boyd, with input from Yinyu Ye; and was implemented
by Michael Grant [GBY06]. It incorporates ideas from earlier works by Löfberg [Löf04], Dahl and [DV04],
Wu and Boyd [WB00], and many others. The modeling language follows the spirit of AMPL or GAMS;
unlike these packages, however, CVX was designed from the beginning to fully exploit convexity. The
specific method for implementing CVX in Matlab draws heavily from YALMIP.
We wish to thank the following people for their contributions: Toh Kim Chuan, Laurent El Ghaoui, Arpita
Ghosh, Siddharth Joshi, Johan Löberg, Almir Mutapcic, Michael Overton and his students, Art Owen, Rahul
Panicker, Imre Polik, Joëlle Skaf, Lieven Vandenberghe, Argyris Zymnis. We are also grateful to the many
students in several universities who have (perhaps unwittingly) served as beta testers by using CVX in their
classwork. We thank Igal Sason for catching many typos in an earlier version of this document, and generally
helping us to improve its clarity.
We would like to thank Gurobi Optimization and MOSEK ApS for their generous assistance as we developed
the interfaces to their commercial products.
83
The CVX Users’ Guide, Release 2.1
84
Chapter 14. Credits and Acknowledgements
CHAPTER
FIFTEEN
USING GUROBI WITH CVX
15.1 About Gurobi
Gurobi Optimization was founded in 2008 by some of the most experienced and respected members of
the optimization community. The Gurobi solver quickly became an industry performance leader in linear,
quadratic, and mixed-integer programming. Gurobi is a fantastic solver for use with CVX, particularly with
the new integer and binary variable capability added in CVX 2.0.
Using CVX with Gurobi requires both a CVX Professional license and a Gurobi license. Please visit Licensing for information about CVX licensing, and Gurobi’s Licensing Overview page for information about
Gurobi licenses.
Academic users can obtain both licenses at no charge. An academic CVX Professional license is obtained
by submitting an Academic License Request. For instructions on obtaining an academic license for Gurobi,
please see Gurobi’s Academic Licenses page.
15.2 Using the bundled version of Gurobi
Note: The bundled version of Gurobi can only be used within CVX. If you wish to use Gurobi outside of
CVX as well, you will need a standalone Gurobi installation.
If you wish to use CVX with the bundled version of Gurobi, you will need three things:
• A CVX Professional Solver Bundle, available here.
• A Gurobi license code, which is composed of 32 hexidecimal digits in the format
xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx. If you purchase a commercial CVX+Gurobi
package, you will receive this code in an email from CVX Research. If you are an academic user, you
will receive it directly from Gurobi.
• A CVX Professional license, saved to a convenient location on your local disk.
Installation proceeds as follows:
• First, install CVX in the standard manner according to the directions found in Installation. Do not
attempt to install either license at this stage.
85
The CVX Users’ Guide, Release 2.1
• Next, retrieve your Gurobi license key by running the command cvx_grbgetkey {code}, where
{code} is the 32-digit Gurobi key. The command will look something like this:
cvx_grbgetkey xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
Important note for academic users: this step must be run from a computer connected to your university
network (a VPN is usually sufficient). Please consult this page of the Gurobi documentation for
details.
• Finally, install your CVX Professional license according to the directions found in Installing a CVX
Professional license.
If you complete these steps successfully, cvx_setup will add Gurobi to its solver list. If you have an
academic or dual-solver CVX Professional license, the MOSEK solver will be added to the solver list as
well. If for some reason installation fails, the output of cvx_setup will provide diagnostic information
that you can use to rectify the problem. If you are still unable to complete the installation, feel free to contact
CVX Support.
15.3 Using CVX with a standalone Gurobi installation
If you wish to use CVX with a standalone installation of Gurobi, you will need the following four things:
• A Gurobi installation package, or a pre-existing Gurobi installation. CVX works with Gurobi 5.0 or
later, but use of the latest version is always recommended.
• A Gurobi license code or key file, if you are installing Gurobi for the first time.
• A standard CVX package, available here. You do not need the Professional Solver bundle.
• A CVX Professional license, saved to a convenient location on your local disk.
Installation proceeds as follows:
• Install Gurobi. See Downloading and Installation from the Gurobi Quick Start Guide.
• Install the Gurobi license. See How to Obtain and Install a Gurobi License.
• Connect your Gurobi installation to MATLAB. See Setting up Gurobi for MATLAB.
• Install CVX and/or the CVX Professional license. See Installation and Installing a CVX Professional
license. Even if these have already been installed, you must at least re-run cvx_setup so that CVX
can locate Gurobi and add it to your solver list.
If you complete these steps successfully, cvx_setup will show that Gurobi has been recognized and added
to the solver list. If for some reason installation fails, the output of cvx_setup will provide diagnostic
information that you can use to rectify the problem. If you are still unable to complete the installation, feel
free to contact CVX Support.
15.4 Selecting Gurobi as your default solver
Even if Gurobi is successfully added to your solver list, it will not automatically be selected as your default
solver. To change this, type the following two commands on the MATLAB command line:
86
Chapter 15. Using Gurobi with CVX
The CVX Users’ Guide, Release 2.1
cvx_solver gurobi
cvx_save_prefs
The first command changes the active solver to Gurobi, but only for the current session. The second line
saves that change to CVX’s preference file, so that Gurobi will be selected as the active solver every time
you start MATLAB.
15.5 Obtaining support for CVX and Gurobi
If you encounter problems using CVX and Gurobi, please contact CVX Support first instead of Gurobi
Optimization. If we can reproduce your problem, we will determine whether or not it is an issue that is
unique to CVX or needs to be forwarded to Gurobi for further analysis.
15.5. Obtaining support for CVX and Gurobi
87
The CVX Users’ Guide, Release 2.1
88
Chapter 15. Using Gurobi with CVX
CHAPTER
SIXTEEN
USING MOSEK WITH CVX
16.1 About MOSEK
MOSEK ApS is widely considered the leader in commercial software for nonlinear convex optimization.
The company is led by CEO Erling Andersen, and its board is chaired by Stanford Professor Yinyu Ye. Both
are internationally recognized for their contributions to the field of convex optimization, and remain active
in research and publication. With its existing support for integer variables and the addition of semidefinite
programming capability in version 7.0, the MOSEK solver can address a wider variety of CVX models than
any other solver.
Using CVX with MOSEK requires a CVX Professional license. Please visit Licensing to learn more about
licensing options. Academic users can obtain a free CVX Professional license by submitting an Academic
License Request.
Note: If you intend to use CVX with both MOSEK and Gurobi, please follow the directions on the page
Using Gurobi with CVX.
16.2 Using the bundled version of MOSEK
Note: The bundled version of MOSEK can only be used within CVX. If you wish to use MOSEK outside
of CVX as well, you will need a standalone MOSEK installation.
The simplest way to use MOSEK with CVX is by installing the appropriate CVX Professional Solver Bundle, available here, along with a MOSEK-enabled CVX Professional license. Please see Installation and
Installing a CVX Professional license for general installation instructions. Once the CVX Professional license has been properly installed, MOSEK will be enabled.
16.3 Using CVX with separate MOSEK installation
If you wish to use CVX with a separate installation of MOSEK 6.0 or 7.0, follow these steps after you have
successfully installed MOSEK:
89
The CVX Users’ Guide, Release 2.1
• Make sure that MATLAB can locate your current installation of MOSEK. If you have already been
using the mosekopt command within MATLAB, there is no further configuration needed. Otherwise, you will need to modify your MATLAB search path so it can find your MOSEK installation.
For information, please see the relevant page in your MOSEK documentation:
– MOSEK 7: Installation
– MOSEK 6: Insatllation
• If you have not done so yet, download and install CVX and a CVX Professional license according to
the instructions in Installation and Installing a CVX Professional license.
• If CVX and your license had already been installed, simply re-run cvx_setup so that CVX can
rebuild its solver list and include MOSEK.
If successful, the output of cvx_setup should show that MOSEK is among the list of available solvers. If
it fails to find MOSEK, it will offer diagnostic information that you can use to correct the problem. If those
remedies fail, feel free to contact CVX Support.
16.4 Selecting MOSEK as your default solver
Even if MOSEK is successfully added to your solver list, it will not automatically be selected as your default
solver. To change this, type the following two commands on the MATLAB command line:
cvx_solver mosek
cvx_save_prefs
The first command changes the active solver to MOSEK, but only for the current session. The second line
saves that change to CVX’s preference file, so that MOSEK will be selected as the active solver every time
you start MATLAB.
16.5 Obtaining support for CVX and MOSEK
If you encounter problems using CVX and MOSEK, please contact CVX Support first instead of MOSEK
ApS. If we can reproduce your problem, we will determine whether or not it is an issue that is unique to
CVX or needs to be forwarded to MOSEK ApS for further analysis.
90
Chapter 16. Using MOSEK with CVX
BIBLIOGRAPHY
[AG00] F. Alizadeh and D. Goldfarb. Second-order cone programming. Mathematical Programming, Series
B, 95:3-51, 2001. http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.23.5133
[BKVH05] S. Boyd, S. J. Kim, L. Vandenberghe, and A. Hassibi,. A tutorial on geometric programming.
Optimization and Engineering, 8(1):67-127, 2007. http://stanford.edu/~boyd/papers/gp_tutorial.html
[BV04] S. Boyd and L. Vandenberghe. Convex Optimization. Cambridge University Press, 2004.
http://stanford.edu/~boyd/cvxbook.html
[Cru02] C. Crusius. A Parser/Solver for Convex Optimization Problems. Ph.D. thesis, Information Systems
Laboratory, Department of Electrical Engineering, Stanford University, 2002.
[DV04] J. Dahl and L. Vandenberghe, CVXOPT: A Python package for convex optimization (version 1.1.5).
http://abel.ee.ucla.edu/cvxopt/
[GBY06] M. Grant and S. Boyd and Y. Ye. Disciplined convex programming. In Global Optimization: from
Theory to Implementation, Nonconvex Optimization and Its Applications, L. Liberti and N. Maculan,
eds., Springer, 2006. http://stanford.edu/~boyd/disc_cvx_prog.html
[Gra04] M. Grant. Disciplined Convex Programming. Ph.D. thesis,
tems Laboratory, Department of Electrical Engineering, Stanford
http://stanford.edu/~boyd/disc_cvx_prog.html
Information SysUniversity, 2004.
[Löf04] J. Löfberg. YALMIP: a toolbox for modeling and optimization in MATLAB. Proceedings of the
2004 International Symposium on Computer Aided Control Systems Design, IEEE Press, September
2004, pp. 284-289. http://users.isy.liu.se/johanl/yalmip/
[Owen06] A. Owen. A robust hybrid of lasso and ridge regression. Technical report, Department of Statistics, Stanford University, October 2006. http://www-stat.stanford.edu/~owen/reports/hhu.pdf
[Stu99] J.F. Sturm, Using SeDuMi 1.02, a MATLAB toolbox for optimization over symmetric cones. Optimization Methods and Software, 11-12:625-633, 1999. Special issue on Interior Point Methods (CD
supplement with software). http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.49.6954
[TTT03] R.H. Tütüncü, K.C. Toh, and M.J. Todd. Solving semidefinite-quadratic-linear
programs using SDPT3. Mathematical Programming, Series B, 95:189-217, 2003.
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.127.4807
91
The CVX Users’ Guide, Release 2.1
[WB00] S.P. Wu and S. Boyd. SDPSOL: A parser/solver for semidefinite programs with matrix structure. In
Recent Advances in LMI Methods for Control, L. El Ghaoui and S.I. Niculescu, eds., SIAM, pp. 79-91,
2000. http://www.stanford.edu/~boyd/sdpsol.html
92
Bibliography
INDEX
A
Academic licensing, 3
B
Bound-constrained least squares, 11
C
CVX, 1
installing, 4
CVX Professional, 3
cvx_setup, 5
D
DCP, 1, 2
DCP ruleset, 2
Disciplined convex program, see DCP
E
linprog (MATLAB function), 12
Linux, 5
LP, 1
M
Mac, 5
Matlab, 1
versions, 5
MIDCP, 1, 2
Mixed-integer disciplined convex program, see
MIDCP
Mixed-integer problems, see MIDCP
MOSEK, 1, 6
P
Platforms, 5
Q
Examples, 7
QP, 1
Quadratic program, see QP
G
S
Geometric program, see GP
GP, 1
GP mode, 1
Gurobi, 1, 6
SDP, 1
SDP mode, 1
SDPT3, 1, 6
Second-order cone program, see SOCP
SeDuMi, 1, 6
Semidefinite program, see SDP
SOCP, 1
Solvers, 1, 6
commercial, 1
Gurobi, 1, 6
included, 6
MOSEK, 1, 6
SDPT3, 1, 6
SeDuMi, 1, 6
I
Installation, 4
L
Least squares, 9
bound-constrained, 11
License, 3
academic, 3
commercial, 3
free, 3
installing, 6
Linear program, see LP
W
Windows, 5
93
Source Exif Data:
File Type : PDF File Type Extension : pdf MIME Type : application/pdf PDF Version : 1.4 Linearized : No Page Count : 99 Page Mode : UseOutlines Warning : Duplicate 'Author' entry in dictionary (ignored) Author : Michael C. Grant, Stephen P. Boyd, CVX Research, Inc. Title : The CVX Users' Guide Subject : Creator : LaTeX with hyperref package Producer : pdfTeX-1.40.3 Create Date : 2015:06:10 21:44:07-04:00 Modify Date : 2015:06:10 21:44:07-04:00 Trapped : False PTEX Fullbanner : This is pdfTeX using libpoppler, Version 3.141592-1.40.3-2.2 (Web2C 7.5.6) kpathsea version 3.5.6EXIF Metadata provided by EXIF.tools