5 The Command Line Tool

5.1 Introduction

The MOSEK command line tool is used to solve optimization problems from the operating system command line. It is invoked as follows

mosek [options] [filename]

where both [options] and [filename] are optional arguments:

  • [options] consists of command line arguments that modify the behavior of MOSEK. They are listed in Sec. 5.5 (Command Line Arguments). In particular, options can be used to set optimizer parameters.

  • [filename] is a file describing the optimization problem. The MOSEK command line accepts files in any of the supported file formats or in the AMPL .nl format.

If no arguments are given, MOSEK will display a splash screen and exit.

user@host:~$ mosek/8/tools/platform/linux64x86/bin/mosek

MOSEK Version 8.0.0.32(BETA) (Build date: 2016-7-12 10:29:26)
Copyright (c) MOSEK ApS, Denmark. WWW: mosek.com
Platform: Linux/64-X86


*** No input file specified. No optimization is performed.

Return code - 0  [MSK_RES_OK]

5.2 Files

The MOSEK command line tool communicates with the user via files and prints some execution logs and solution summary to the terminal.

Input files

Optimization problems are read from files. See Sec. 11 (Supported File Formats) for details.

File format conversion

To convert between two file formats supported by MOSEK use the option -x together with -out to specify the target file name. The target file type must support the problem type of the source file, otherwise the conversion will be partial. For instance in case a MPS file must be converted in a more readable OPF format, the following line can be used

mosek -x -out lo1.opf lo1.mps

With the -x option the solver will not actually solve the problem.

Output files

Solutions are written to files:

  • .bas - basic solution,

  • .sol - interior point solution,

  • .itg - integer solution (the only available solution for mixed-integer problems).

For linear problems both the basic and interior point solution may be present. Infeasibility certificates are stored in the same files. See Sec. 11.8 (The Solution File Format) for details.

5.3 Example

To solve a problem stored in file, say lo1.mps, write:

mosek lo1.mps

The solver will

  • read lo1.mps from disk,

  • solve the problem and display the solution log and

  • store the relevant solution files if any solution exists; file content explained in Sec. 11.8 (The Solution File Format).

MOSEK Version 8.0.0.34(BETA) (Build date: 2016-8-24 00:51:13)
Copyright (c) MOSEK ApS, Denmark. WWW: mosek.com
Platform: Linux/64-X86

Open file '/home/andrea/mosek/8/tools/examples/data/lo1.mps'
Reading started.
Using 'obj' as objective vector
Read 13 number of A nonzeros in 0.00 seconds.
Using 'rhs' as rhs vector
Using 'bound' as bound vector
Reading terminated. Time: 0.00

Read summary
  Type             : LO (linear optimization problem)
  Objective sense  : max
  Scalar variables : 4
  Matrix variables : 0
  Constraints      : 3
  Cones            : 0
  Time             : 0.0

Problem
  Name                   : lo1
  Objective sense        : max
  Type                   : LO (linear optimization problem)
  Constraints            : 3
  Cones                  : 0
  Scalar variables       : 4
  Matrix variables       : 0
  Integer variables      : 0

Optimizer started.
Interior-point optimizer started.
Presolve started.
Linear dependency checker started.
Linear dependency checker terminated.
Eliminator started.
Freed constraints in eliminator : 0
Eliminator terminated.
Eliminator - tries                  : 1                 time                   : 0.00
Lin. dep.  - tries                  : 1                 time                   : 0.00
Lin. dep.  - number                 : 0
Presolve terminated. Time: 0.00
Optimizer  - threads                : 2
Optimizer  - solved problem         : the primal
Optimizer  - Constraints            : 3
Optimizer  - Cones                  : 0
Optimizer  - Scalar variables       : 6                 conic                  : 0
Optimizer  - Semi-definite variables: 0                 scalarized             : 0
Factor     - setup time             : 0.00              dense det. time        : 0.00
Factor     - ML order time          : 0.00              GP order time          : 0.00
Factor     - nonzeros before factor : 6                 after factor           : 6
Factor     - dense dim.             : 0                 flops                  : 1.06e+02
ITE PFEAS    DFEAS    GFEAS    PRSTATUS   POBJ              DOBJ              MU       TIME
0   8.0e+00  3.2e+00  3.5e+00  1.00e+00   1.000000000e+01   0.000000000e+00   1.0e+00  0.01
1   4.2e+00  2.5e+00  4.7e-01  0.00e+00   3.093970927e+01   2.766058702e+01   2.6e+00  0.01
2   4.2e-01  2.5e-01  4.6e-02  -1.82e-02  6.511676243e+01   6.308843559e+01   2.6e-01  0.01
3   3.6e-02  2.1e-02  3.9e-03  5.84e-01   8.096141239e+01   8.061962333e+01   2.2e-02  0.01
4   1.5e-05  9.1e-06  1.7e-06  9.43e-01   8.333280389e+01   8.333241803e+01   9.2e-06  0.01
5   1.5e-09  9.1e-10  1.7e-10  1.00e+00   8.333333328e+01   8.333333324e+01   9.2e-10  0.01
Basis identification started.
Primal basis identification phase started.
ITER      TIME
0         0.00
Primal basis identification phase terminated. Time: 0.00
Dual basis identification phase started.
ITER      TIME
0         0.00
Dual basis identification phase terminated. Time: 0.00
Basis identification terminated. Time: 0.00
Interior-point optimizer terminated. Time: 0.01.

Optimizer terminated. Time: 0.02

Interior-point solution summary
  Problem status  : PRIMAL_AND_DUAL_FEASIBLE
  Solution status : OPTIMAL
  Primal.  obj: 8.3333333280e+01    nrm: 5e+01    Viol.  con: 1e-08    var: 0e+00
  Dual.    obj: 8.3333333242e+01    nrm: 4e+00    Viol.  con: 2e-10    var: 5e-09

Basic solution summary
  Problem status  : PRIMAL_AND_DUAL_FEASIBLE
  Solution status : OPTIMAL
  Primal.  obj: 8.3333333333e+01    nrm: 5e+01    Viol.  con: 7e-15    var: 0e+00
  Dual.    obj: 8.3333333245e+01    nrm: 4e+00    Viol.  con: 2e-10    var: 5e-09

Optimizer summary
  Optimizer                 -                        time: 0.02
    Interior-point          - iterations : 5         time: 0.01
      Basis identification  -                        time: 0.00
        Primal              - iterations : 0         time: 0.00
        Dual                - iterations : 0         time: 0.00
        Clean primal        - iterations : 0         time: 0.00
        Clean dual          - iterations : 0         time: 0.00
    Simplex                 -                        time: 0.00
      Primal simplex        - iterations : 0         time: 0.00
      Dual simplex          - iterations : 0         time: 0.00
    Mixed integer           - relaxations: 0         time: 0.00

Open file '/home/andrea/mosek/8/tools/examples/data/lo1.sol'
Start writing.
done writing. Time: 0.00

Open file '/home/andrea/mosek/8/tools/examples/data/lo1.bas'
Start writing.
done writing. Time: 0.00

Return code - 0  [MSK_RES_OK]

5.4 Solver Parameters

MOSEK comes with a large number of parameters that allows the user to tune the behavior of the optimizer. The typical settings which can be changed with solver parameters include:

  • choice of the optimizer for linear problems,

  • choice of primal/dual solver,

  • turning presolve on/off,

  • turning heuristics in the mixed-integer optimizer on/off,

  • level of multi-threading,

  • feasibility tolerances,

  • solver termination criteria,

  • behaviour of the license manager,

and more. All parameters have default settings which will be suitable for most typical users. Each parameter is identified by a unique string name and it can accept either integers or symbolic names, floating point values or symbolic strings. Please refer to Sec. 10.2 (Parameters (alphabetical list sorted by type)) for the complete list of available solver parameters.

5.4.1 Setting from command line

Setting solver parameters is possible using the command line option -d.If multiple parameters must be specified, option -d must be repeated for each one. For example, the next command will switch off presolve, set a feasibility tolerance and solve the problem from lo1.opf:

mosek -d MSK_IPAR_PRESOLVE_USE MSK_OFF -d MSK_DPAR_INTPNT_TOL_PFEAS 1.0e-8 lo1.opf

5.4.2 Using the Parameter File

Solver parameters can also be set using a parameter file, for example:

BEGIN MOSEK
% This is a comment.
% The subsequent line tells MOSEK that an optimal
% basis should be computed by the interior-point optimizer.
MSK_IPAR_PRESOLVE_USE         MSK_OFF
MSK_DPAR_INTPNT_TOL_PFEAS     1.0e-9
END MOSEK

The syntax of the parameter file must obey a few simple rules:

  • The file must begin with BEGIN MOSEK and end with END MOSEK.

  • Empty lines and lines starting from a % sign are ignored.

  • Each line contains a valid MOSEK parameter name followed by its value.

The parameter file can have any name. Assuming it has been called mosek.par, it can be used using the -p option as follows:

mosek -p mosek.par afiro.mps

Command-line parameters override those from the parameter file in case of repetition. For instance

mosek -p mosek.par -d MSK_DPAR_INTPNT_TOL_PFEAS 1.0e-8 afiro.mps

will set MSK_DPAR_INTPNT_TOL_PFEAS to \(10^{-8}\) using the value provided on the command line.

5.5 Command Line Arguments

The following list shows the available command-line arguments for MOSEK:

-anapro

Analyze the problem data.

-anasoli <name>

Analyze the initial solution name e.g. -anasoli bas.

-anasolo <name>

Analyze the final solution name e.g. -anasolo itg.

-basi <name>

Input basic solution file name.

-baso <name>

Output basic solution file name.

-d  <name>  <value>

Define the value value for the MOSEK parameter name.

-dbgmem  <name>

Name of memory debug file.

-dualout  <name>

Output dual problem file name.

-f

Complete license information is printed.

-h, -?

Help.

-inti <name>

Input integer solution file name.

-into <name>

Output integer solution file name.

-itri <name>

Input interior point solution file name.

-itro <name>

Output interior point solution file name.

-info <name>

Infeasible subproblem output file name.

-infrepo <name>

Feasibility reparation output file.

-jsoli <name>

Input JSON format solution file name.

-jsolo <name>

Output JSON format solution file name.

-l,-L <dir>

dir is the directory where the MOSEK license file mosek.lic is located.

-max

The problem is maximized.

-min

The problem is minimized.

-n

Ignore errors in subsequent parameter settings.

-optserv <url>

Use an OptServer specified by an URL of the form http://host:port.

-out <name>

Write the task to a data file named name. See Sec. 11 (Supported File Formats).

-p <name>, -pari <name>

Name of the input parameter file.

-paro <name>

Name of the output parameter file.

-primalrepair

Repair a primal infeasible problem. See Sec. 9.2 (Automatic Repair of Infeasible Problems).

-r

If the option is present, the program returns \(-1\) if an error occurred, otherwise \(0\).

-removeitg

Removes all integer constraints after reading the problem.

-rout <name>

If the option is present, the program writes the return code to file name.

-q <name>

Name of an optional log file.

-sen <file>

Perform sensitivity analysis based on file.

-silent

As little information as possible is send to the terminal.

-toconic

Translate to conic form after reading.

-v

MOSEK version is printed and no optimization is performed.

-w

If this options is on, then MOSEK will wait for a license.

-x

Do not run the optimizer. Useful for converting between file formats.

-=

List all possible solver parameters with default value, lower bound and upper bound (if applicable).

5.6 The license system

MOSEK is a commercial product that always needs a valid license to work. MOSEK uses a third party license manager to implement license checking. The number of license tokens provided determines the number of optimizations that can be run simultaneously.

By default a license token remains checked out for the whole execution of the command line tool. If the license is not unlimited, then the number of tokens determines the maximal number of processes that can run simultaneously. In this case setting the license wait flag with the parameter MSK_IPAR_LICENSE_WAIT will force MOSEK to wait until a license token becomes available instead of returning with an error.