Mixed integer programming formulations for single machine scheduling problems

In this paper, the computational performance of four different mixed integer programming (MIP) formulations for various single machine scheduling problems is studied. Based on the computational results, we discuss which MIP formulation might work best for these problems. The results also reveal that for certain problems a less frequently used MIP formulation is computationally more efficient in practice than commonly used MIP formulations. We further present two sets of inequalities that can be used to improve the formulation with assignment and positional date variables.