SLPlib: Input Routines and Data Structures for Stochastic Linear Programming


Andy Felt, Jason Sarich, and K. A. Ariyawansa


Brief Description

SLPlib is an open source library for stochastic linear programming. The data structures store the stochastic linear program (SLP) in a multistage tree structure for fast access to the data and easy parallelization. The data matrices are stored in sparse difference matrix structures, to minimize data duplication and memory requirements.

There are also routines that read SMPS data files and store the data in the data structures. An option allows the use of Parallel Virtual Machine (PVM) to distribute child nodes to other machines. If this option is used, the entire SLP is never constructed on the master machine. This allows large problems to be specified---problems that would not fit on any single machine.

Additionally, you can randomly generate two stage SLPs. The distributed memory option is also available with these problems.

Visit the Download Page

You are welcome to freely download the library.

Please Contribute

The library is open source, which means if you don't like it, you can improve it. If you change it, please send your changes to me for possible addition to the library. Here is a partial to do list (and hence a list of limitations):
  1. Allow ranges and bounds in the core file.
  2. Allow scenario mode in the stoch file.
  3. Allow multistage input. This is all set except for the reading of the stoch file.

Other Input Routines

Go To Andy Felt's homepage.

You may find my email address in the following directory.

Page author: Andy Felt

Last modified: 11 June, 2001