An improved multi-objective memetic algorithm for bi-objective permutation flow shop scheduling

Abstract

A permutation flowshop scheduling problem of optimizing the makespan and the total flow time, which can be expressed as Fm|prum|(Cmax, ΣCi), is considered in this paper. An improved multi-objective memetic algorithm (IMOMA) is proposed due to the NP-hardness of the problem. In order to effectively trade-off between two objectives, we propose a NEH and LR heuristic based initialization strategy and a powerful local search strategy, in the searching framework of memetic algorithm. Finally, we perform computational experiments by solving ten largest scale instances of Taillard benchmarks, with 500 jobs and 20 machines. The results demonstrate that the proposed IMOMA outperforms the NEHFF heuristic and two state-of-the-art evolutionary multi-objective algorithms, NSGA-II and MOEA/D with respect to convergence and diversity.