Integer Local Search: WSAT(OIP)

WSAT(OIP) is a domain-independent local search method for linear integer constraints that can solve a range of hard realistic integer optimization problems. WSAT(OIP) subsumes WSAT(PB), a code for pseudo-Boolean constraints.

WSAT(OIP) has been demonstrated to be effective on large-scale constraint and integer optimization problems. Case studies have been carried out on capacitated production planning, airport gate allocation, sports scheduling (ACC Basketball League 97/98), radar surveillance (covering-type problems), course assignment, the Progressive Party Problem. Publications of the algorithm and of the case studies are available from this page, including two publications at AAAI97 and AAAI98.

 WSAT(OIP) V1.105 interfaces with AMPL, a modelling language for mathematical programming and the CPLEX callable library. WSAT(OIP) additionally supports inequalities over non-binary variables (finite domain variables).

WSAT(OIP) is available free of charge on a limited variety of platforms, for non-commercial purposes. (Currently, only the Linux version has AMPL/CPLEX enabled.) The code is written in C/C++. The compressed tar archive contains the WSAT(OIP) binary, a dynamic link library to interface with Oz, and an AMPL demo. WSAT(OIP) can read an extension of standard lp_solve (.opb) format for soft and hard constraints (.oip), or alternatively, problems can be modelled using the AMPL or Oz interfaces.
It would be kind if you could complete the following form in order to ftp WSAT(OIP). In return, we will e-mail you information when a new version of the solver becomes available (no junk-mail!).
Please enter your name: 
and your e-mail address: 
Please send questions or bug reports to walser@ps.uni-sb.de (highly appreciated!). 


This page maintained by: J.P. Walser

Last Updated: Dec 2, 1997