As of 2001, researchers at the University of Tilburg (CentER Applied Research) have developed a new combined set partitioning and set packing solver based on Constraint Logic Programming concepts, Lagrangian Relaxation bounds, and newly discovered pre-solve techniques. This research solver does not require the presence of a linear programming code, and is designed to solve large model instances.
The performance of the code is comparable to that of the best mixed-integer programming codes, and is exceptionally good for model instances in which feasibility and/or multiple solutions play an essential role. This solver has been linked to AIMMS, which has benefited both the development of the solver and its subsequent application to a variety of problems.
Anyone interested in experimenting with this solver may contact one of its authors, Prof.dr. H.A. Fleuren, by sending an email to Fleuren@uvt.nl.
Associated publications by M.G.C. van Krieken, H.A. Fleuren, M.J.P. Peeters:
- A Lagrangean relaxation based algorithm for solving Set Partitioning Problems
- Problem reduction in Set Partitioning Problems
Important: This solver is not actively supported.