16.01.2015 Views

GAMS — The Solver Manuals - Available Software

GAMS — The Solver Manuals - Available Software

GAMS — The Solver Manuals - Available Software

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

48 COIN-OR<br />

also found their way into the regular <strong>GAMS</strong> distribution. <strong>The</strong>y are currently available for Linux (32 and 64 bit),<br />

Windows (32 and 64 bit), Sun Solaris (Intel 64 bit), and Darwin (Intel 32 bit) systems. With the availability<br />

of source code for the <strong>GAMS</strong>links the user is not limited to the out of the box solvers that come with a regular<br />

<strong>GAMS</strong> distribution, but can extend and build these interfaces by themselves.<br />

As of <strong>GAMS</strong> Distribution 22.8, available solvers include:<br />

• CoinBonmin 0.99: Basic Open-source Nonlinear Mixed Integer programming<br />

(model types: LP, RMIP, MIP, DNLP, NLP, RMINLP, MINLP, QCP, RMIQCP, MIQCP)<br />

• CoinCbc 2.1: COIN-OR Branch and Cut<br />

(model types: LP, MIP, RMIP)<br />

• CoinGlpk 4.26: Gnu Linear Programming Kit<br />

(model types: LP, MIP, RMIP)<br />

• CoinIpopt 3.4: Interior Point Optimizer<br />

(model types: LP, RMIP, DNLP, NLP, RMINLP, QCP, RMIQCP)<br />

• CoinScip 1.00: Solving Constrained Integer Programs<br />

(model types: LP, MIP, RMIP)<br />

For more information see the COIN-OR/<strong>GAMS</strong>links web site at https://projects.coin-or.org/<strong>GAMS</strong>links.<br />

2 CoinBonmin<br />

<strong>GAMS</strong>/CoinBonmin brings the open source MINLP solver Bonmin from the COIN-OR foundation to the broad<br />

audience of <strong>GAMS</strong> users.<br />

Bonmin (Basic Open-source Nonlinear Mixed Integer programming) 0.9 is an open-source solver for mixedinteger<br />

nonlinear programming (MINLPs), whereof some parts are still experimental. <strong>The</strong> code is developed in<br />

a joined project of IBM and the Carnegie Mellon University. <strong>The</strong> COIN-OR project leader for Bonmin is Pierre<br />

Bonami.<br />

Bonmin implements five different algorithms:<br />

• B-BB: a simple branch-and-bound algorithm based on solving a continuous nonlinear program at each node<br />

of the search tree and branching on variables (default)<br />

• B-OA: an outer-approximation based decomposition algorithm<br />

• B-QG: an outer-approximation based branch-and-cut algorithm (by Queseda and Grossmann)<br />

• B-Hyb: a hybrid outer-approximation/nonlinear programming based branch-and-cut algorithm<br />

• B-Ecp: an ECP cuts based branch-and-cut algorithm a la FilMINT<br />

<strong>The</strong> algorithms are exact when the problem is convex, otherwise they are heuristics.<br />

For convex MINLPs, experiments on a reasonably large test set of problems have shown that B-Hyp is the<br />

algorithm of choice (it solved most of the problems in 3 hours of computing time). Nevertheless, there are cases<br />

where B-OA is much faster than B-Hyb and others where B-BB is interesting. B-QG corresponds mainly to a<br />

specific parameter setting of B-Hyb where some features are disabled. B-Ecp has just been added to Bonmin and<br />

is still experimental. For nonconvex MINLPs, it is strongly recommend to use B-BB, even though it is only a<br />

heuristic for such problems too.<br />

For more information we refer to<br />

• the Bonmin web site https://projects.coin-or.org/Bonmin and

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!