Optimization Packages Available on RCS

AMPL
AMPL is a mathematical programming and optimization modeling language. You can input your model into AMPL in a reasonably intuitive way and it will use a solver (such as MINOS or CPLEX) for solving the problem. It is capable of solving linear, nonlinear, and integer programs. There is a writeup for using AMPL on the RCS system.
Examples from the AMPL website. (There are two RPI alumni mentioned on this page: Bob Bosch and Bob Vanderbei.)
The main AMPL WWW page is at Bell Labs. You can download the first chapter of the book, and the site also has other introductory material available, as well as letting you submit a problem for solving.
Here is more information about AMPL.
CPLEX
CPLEX is a solver for linear programming and integer programming problems. It contains both simplex and interior point algorithms for linear programming problems. It can be used directly, or it can be used as a subroutine. To set up your account to use the student version of CPLEX, you need to modify your .login_profile file as described here. Here is a pdf file containing the AMPL CPLEX 9.0 User Guide. The CPLEX distribution of AMPL examples can be found here. An online manual is available for CPLEX 6.5. CPLEX has a WWW page. An input file to CPLEX should be in MPS format.
MATLAB
MATLAB is an interactive, matrix-based system for scientific and engineering calculations. You can solve numerical problems without having to write a program, although you can also write programs in MATLAB. MATLAB has a WWW page. There is a primer available on using MATLAB.
MOSEK
MOSEK solves similar problems to CPLEX. In addition, it has a state-of-the-art customized solver for conic optimization problems. MOSEK is not installed on RCS. However, MOSEK offers a free full-blown version to students, and is competitive in performance with CPLEX. It can interface with both AMPL and MATLAB.
The NEOS Server:
As an alternative to running your problem locally, you can run it remotely on a computer at Argonne National Lab. The NEOS Server is a project to make the latest techniques in optimization widely available. You can submit a problem in an appropriate framework, and request that a particular algorithm be used to solve the problem. This is a service of the Optimization Technology Center at Argonne National Lab. It includes interactive formulations of the diet problem and the portfolio optimization problem.

Other Packages:
I have an unorganized list of announcements of computational packages that are available for solving a variety of problems. Very few of these have been installed on campus, as far as I know.

Frequently asked questions about AFS.

Documentation on how to use the NIC cluster.

Documentation of the performance of various computers, due to Jack Dongarra

RPI Math


Back to John Mitchell's homepage