Effect of Varying Levels of Setup Time on Makespan Criteria: A Comparative Study of Two Algorithms on SDST Flow Shop

Abstract

Scheduling problems with setup time consideration are the realistic versions of classical scheduling problems. This paper analyses the effect of setup time level variation on the performance of two constructive algorithms for flow shop scheduling with sequence dependent setup time. The two constructive algorithms considered in the study differ in their approach in constructing the sequence; the first algorithm is processing time based while the second algorithm is setup time and processing time based. The setup time based algorithm known as fictitious job setup ranking algorithm is developed as a part of the present study. Taillard benchmark flow shop scheduling problems are used for development of benchmark problems for flow shop scheduling with sequence dependent setup time. The two algorithms are tested for the set of benchmark problem instances developed. The results are analyzed graphically and statistically for the varying levels of setup time. The study reveals that the setup time based algorithm improves its performance when the setup time increases as percentage of processing time. This paper proposes a framework namely setup dominance matrix, which serves as an aid to identify the algorithm that is suitable for a specific size flow shop scheduling problem at a particular level of setup time.