An OBL Harmony Search for Hybrid Flow Shop Scheduling with Multiprocessor Tasks Problem

Abstract

This paper develops a new harmony search (HS) algorithm for the hybrid flow shop scheduling problem with multiprocessor tasks, which is known to be NP-hard in the strong sense. The goal is to minimize the makespan. Thus, the proposed HS starts with the application of an opposition based learning (OBL) method to increase the diversity level of the initial harmony memory (HM). Furthermore, we use novel improvisation rules of producing new harmonies based on crossover and mutation operators. Also, neighborhood techniques are employed to enhance the searching and improve the solution quality. The proposed HS is compared with three algorithms from the literature. The computational experiments demonstrate the effectiveness of the proposed HS algorithm in terms of makespan.