ABSTRACT
A considerable amount of literature suggests solving scheduling problems by stating them as integer programs. These recast problems may be solved by standard algorithms, which have been developed for solving general mathematical problems. Translating back, we obtain optimal schedules. This sounds like a very promising approach, but in practice it has not proven to be. The standard mathematical programming algorithms are practically applicable to quite large problems; the recast practical scheduling problems can be substantially larger. Thus the recasting is just that: the inherent difficulties are rephrased, but not into a more tractable form. Empirically this confirms that scheduling problems are in general very difficult and do not just appear to be so.
