mexIPOPT

by Enrico Bertolazzi

This is my rewrite of Peter Carbonetto MATLAB interface for IPOPT a software package for large-scale ​nonlinear optimization.

Source code and documentation for IPOPT can be downloaded from:

https://projects.coin-or.org/Ipopt

On OSX IPOPT can be installed using Homebrew (http://brew.sh).

Why do a job already done?

The original MATLAB interface (https://projects.coin-or.org/Ipopt/wiki/MatlabInterface) seems not maintained for OSX and do not work with recent MATLAB distributions. I have reorganized and simplified the original interface (only internally) and eliminated the bug (due to MATLAB change in the managing of sparse pattern) in the interfacing of sparse pattern between MATLAB and IPOPT. Moreover I tried to improve error catching.

How to compile

If you have an OSX or a LINUX 64 bit OS try to use the corresponding precompiled mex file, otherwise try to compile the mex interface following the instructions. Compilation is not an easy task, I try to be as clear as possibile, in any case things may change from different version of OS.

The interface is written in C++. To compile the mex file

• change the working directory to the lib directory.
• run the script compile_osx or compile_linux or compile_win.

To compile the mex interface you need a static IPOPT library installed in your system. To install the library read README_osx_how_to_compile_ipopt.md or README_linux_how_to_compile_ipopt.md. For windows system in the binary directory I copied precompiled dll of ipopt + dependecties taken from https://github.com/JuliaOpt.

For the usage of the mex file use addpath to add lib directory in the search path or move the contents of lib to a directory in the search path.

Examples

In the directory example you find the original examples of Peter Carbonetto which shows the usage of the interface.

Author:

Enrico Bertolazzi
Department of Industrial Engineering
University of Trento
enrico.bertolazzi@unitn.it