A hybrid MIP/CP approach for multi-activity shift scheduling

Abstract

We propose a hybrid MIP/CP approach for solving multi-activity shift scheduling problems, based on regular languages that partially describe the set of feasible shifts. We use an aggregated MIP relaxation to capture the optimization part of the problem and to get rid of symmetry. Whenever the MIP solver generates a integer solution, we use a CP solver to check whether it can be turned into a feasible solution of the original problem. A MIP-based heuristic is also developed. Computational results are reported, showing that the proposed method is a promising alternative compared to the state-of-the-art. © 2012 Springer-Verlag.