Third-Party Products for Use with GPOPS

NLP Solvers For Use with GPOPS:

GPOPS users now have the choice of two NLP solvers: SNOPT and IPOPT. The directions for obtaining each NLP solver is given below.

OBTAINING THE NLP SOLVER SNOPT FOR USE WITH GPOPS

By a gracious arrangement with Professor Philip Gill of UC, San Diego, non-commercial users of GPOPS users can obtain the SNOPT mex files for SNOPT as follows:

  • Go to the following URL: http://www.cam.ucsd.edu/~peg/gpops/Downloads.html
  • In the username field, type “rao”. In the password field type “anilvdol”
  • Download either the file “snoptMatlabMex.zip” or “snoptMatlabMex.tgz”
  • Uncompress the archive
  • From within MATLAB, change to the directory “snoptMatlabMex”
  • Run the file “snoptSetup.m” from within MATLAB
If you are a commercial organization, then you must purchase a license for SNOPT either from Stanford Business Software, Inc. (http://www.s-b-s-i.com/) or The Office of Technology Licensing at Stanford University (http://otl.stanford.edu/flash.html)

OBTAINING THE NLP SOLVER IPOPT FOR USE WITH GPOPS

IPOPT is an open-source and freely available NLP solver. IPOPT is part of the Computational Infrastructure for Operations Research (COIN-OR) repository. The main website for IPOPT is https://projects.coin-or.org/Ipopt. Extensive instructions for downloading, installing and using IPOPT can be found at http://www.coin-or.org/Ipopt/documentation/. In addition, information on compiling the Matlab interface for IPOPT can be found at https://projects.coin-or.org/Ipopt/wiki/MatlabInterface.


Automatic Differentiators for Use with GPOPS:

While GPOPS now has a built-in automatic differentiator to compute the gradient of the objective function and the Jacobian of the constraints, the user may also choose to use the automatic differentiation packages INTLAB and MAD.

OBTAINING THE AUTOMATIC DIFFERENTIATION PACKAGE INTLAB

Information about INTLAB can be found at the following URL: http://www.ti3.tu-harburg.de/rump/intlab/ (personal or non-commercial users can download INTLAB by scrolling to the bottom of the webpage http://www.ti3.tu-harburg.de/rump/intlab/ and search for their operating system [Unix or Windows]). For commercial or other use, a license for INTLAB is required and can be obtained by contacting Professor Siegfried Rump at the following e-mail address: rump@tu-harburg.de.

OBTAINING THE AUTOMATIC DIFFERENTIATION PACKAGE MAD

MAD (Matlab Automatic Differentiation) is joint development between Cranfield University (UK) and Tomlab Optimization, Inc. Information about MAD can be obtained by visiting the following URL: http://matlabad.com/.