Cutting Big M Down to Size
Published Online:1 Oct 1990https://doi.org/10.1287/inte.20.5.61
Abstract
Some texts and applied research papers continue to advocate the use of “M,” an “extremely large number,” in modeling a fixed charge in binary integer programs. This choice of M can be very detrimental to solving even reasonably sized problems. Previous research provides straightforward techniques which should be implemented by any person intelligent enough to develop an MIP model.
