Heuristics for parallel machine scheduling problem with history dependent setup times

Abstract

This thesis focuses on a parallel-machine scheduling problem for jobs with sequence-independent family setup times and history dependent setup times, which can be found often in semiconductor wafer fabrication facilities. Since the parallel-machine scheduling problem (with the objective of minimizing the total weighted flow time of the jobs) is known to be NP-complete, we present heuristic algorithms, in which a simple and systematic scheduling logic is employed. The scheduling logic is related with decisions and information involved in the scheduling problem, which are job selection and production history generation. We propose rules for production history generation and job selection rules including simple priority rules. A series of computational experiments was performed to compare the suggested heuristics with a scheduling method currently used in practice, and results show that the suggested heuristics outperform the existing method.