Scheduling deteriorating jobs and module changes with incompatible job families on parallel machines using a hybrid SADE-AFSA algorithm

Abstract

This research is motivated by a scheduling problem found in the special steel industry of continuous casting processing, where the special steel is produced on the parallel machines, i.e., the continuous casting machine, and each machine can produce more than one types of special steel. Usually, different types of special steel have diversity alloy content, which generates distinct cooling requirements. Consequently, the job families are incompatible, different types of special steel cannot be continuous process. This indicates that the machine will pause for a period of time to execute the module change activity between two adjacent job families. In this context, we attempt to investigate a parallel machine scheduling problem with the objective of minimizing the makespan, i.e., the completion time of the last job. The effect of deterioration, incompatible job families, and the module change activity are taken into consideration simultaneously, and the actual processing time of each job depends on its starting time and normal processing time. A hybrid SADE-AFSA algorithm combining Self-Adaptive Differential Evolution (SADE) and Artificial fish swarm algorithm (AFSA) is proposed to tackle this problem. Finally, the computational experiments are conducted to evaluate the performance of the proposed algorithm.