A modified adaptive switching-based many-objective evolutionary algorithm for distributed heterogeneous flowshop scheduling with lot-streaming

Abstract

The distributed heterogeneous permutation flowshop scheduling problem with lot-streaming (DHPFSPLS) is a provocative scheduling and optimization problem confronting both industry and engineering. However, no result is available in investigating the DHPFSPLS with variable number of sublots. This paper presents a many-objective mathematical model of this problem with the objectives of makespan, idle time of machines, total production cost and total flow time, considering the transfer time and sequence-independent setup time. Based on this model, a modified adaptive switching-based many-objective evolutionary algorithm is proposed, in which each solution is coded using a three-vector-based solution representation, i.e., a factory assignment vector, a lot-splitting vector and a job permutation vector. Then, a novel two-population collaborative search strategy based on a learning mechanism is designed, which can enhance exploitation abilities and make effective use of optimization knowledge from the population. Moreover, an adaptive switching strategy-based environmental selection is implemented to ensure the convergence and diversity of the solution set. Through a variety of computational tests and comparisons, the effectiveness of the proposed algorithm in solving the many-objective DHPFSPLS is demonstrated.