Mathematical Programming vs. Constraint Programming for Scheduling Problems

Abstract

This paper focuses on a classical scheduling problem known as the job-shop scheduling problem which is one of the most difficult problems in combinatorial optimisation. The paper presents two solution techniques, namely mathematical programming and constraint programming and compares their computational efficiency on benchmark problems. In addition, the experience with scheduling trains in a passenger railway station is presented. The computational experiments proved that the mathematical programming approach outperforms constraint programming with respect to the quality of the solution.