Information about AMPL

Background information and reading

AMPL is a mathematical programming and optimization modeling language which can be used to solve linear, integer, and nonlinear programming problems. 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.

The AMPL website contains the first chapter of the AMPL book, a collection of frequently asked questions, and a list of all the cplex options available in AMPL.

The student version of AMPL

The AMPL website contains downloadable executables for use in Windows or in Unix/Linux systems, including Mac OSX. These executables are limited in the size of problems they can solve.

Note that If you use the command line interface with Windows and you use Word to create and edit the files, Word will add txt to the end of a filename when stored as a text file. For example, if your model is called zoff.mod and you created it and stored it using Word, the file will be called zoff.mod.txt. To read it into AMPL, you will need to use the command

ampl: model zoff.mod.txt;

Size-unrestricted versions of AMPL and CPLEX

You must have an RPI account to be able to use the size-unrestricted versions.

Installing AMPL and CPLEX on your Windows XP or Linux computer: Our license now allows you to install the size-unrestricted versions of these programs on your own machine. Once you've installed the programs in Windows XP, you run them from a command prompt window. In Linux, they run in a terminal window. The program OPL is also available, as an alternative to AMPL. OPL is an ILOG modelling language and environment that has a user interface that is more user-friendly than AMPL. You can even run them from off-campus as long as you have Cisco VPN running.

To use the size-unrestricted version of AMPL on RCS, you need to follow these instructions.

Note: Our license server was changed on March 20, 2006. Therefore, you may need to update your access.ilm file. Instructions for getting this file are in Steps 2 and 3 of the instructions for installing the programs on your machine.

More on solvers

AMPL is a modelling language. It uses other packages to actually solve optimization problems. The default solver we have available on campus is CPLEX. Our current version of CPLEX will solve integer, linear, and convex quadratic programming problems. As of June 15, 2006, it can solve quadratically constrained problems, including quadratically constrained integer programs.

Here is some more information about working with AMPL.

The NEOS Server

AMPL can also be run remotely on the NEOS server. You submit your model and data files and it will give you the solution.

Examples:

Examples from the AMPL website. (There are two RPI alumni mentioned on this page: Bob Bosch and Bob Vanderbei.)
Here are some more AMPL examples.
More examples, including a set formulation of a general production problem together with three data files.
The CPLEX distribution of AMPL examples can be found here.
An excellent collection of nonlinear programming models. A good place to start may be the block QP models in the Cute set, for example blockqp1.mod.

Web resources and hints

More information about working with AMPL and getting more from it.
The CPLEX website.
Information on using the CPLEX callable library, including a sample makefile.
Here is the cplex documentation. This page has links to an overview of the callable library, and also links to material on getting started with cplex and to the user's manual. The user guide is also available as a pdf file, and the remaining documentation can also be downloaded in pdf format.
Debugging an infeasible model by Gabor Pataki. This uses the IIS options within CPLEX. In particular, you can use the AMPL command option cplex_options 'iisfind 1';
Displaying the values of all variables simultaneously.

John Mitchell