Discrete evolutionary multi-objective optimization for energy-efficient blocking flow shop scheduling with setup time

• Published in: Applied soft computing Vol. 93; p. 106343
• Main Authors: Han, Yuyan, Li, Junqing, Sang, Hongyan, Liu, Yiping, Gao, Kaizhou, Pan, Quanke
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.08.2020
• Subjects: Blocking flow shop; Energy consumption; Multi-objective evolutionary optimization; Self-adaptive; Self-adaptive; Energy consumption; Blocking flow shop; Multi-objective evolutionary optimization
• ISSN: 1568-4946; 1872-9681
• DOI: 10.1016/j.asoc.2020.106343

Sustainable scheduling problems have been attracted great attention from researchers. For the flow shop scheduling problems, researches mainly focus on reducing economic costs, and the energy consumption has not yet been well studied up to date especially in the blocking flow shop scheduling problem. Thus, we construct a multi-objective optimization model of the blocking flow shop scheduling problem with makespan and energy consumption criteria. Then a discrete evolutionary multi-objective optimization (DEMO) algorithm is proposed. The three contributions of DEMO are as follows. First, a variable single-objective heuristic is proposed to initialize the population. Second, the self-adaptive exploitation evolution and self-adaptive exploration evolution operators are proposed respectively to obtain high quality solutions. Third, a penalty-based boundary interstation based on the local search, called by PBI-based-local search, is designed to further improve the exploitation capability of the algorithm. Simulation results show that DEMO outperforms the three state-of-the-art algorithms with respect to hypervolume, coverage rate and distance metrics. [Display omitted] •A multi-objective model of BFS scheduling problem with makespan and energy consumption criteria is proposed.•A variant of single-heuristic is incorporated in population initialization.•Two self-adaptive exploitation and exploration evolution strategies are proposed respectively.•A PBI-based-local search is adopted to enhance the exploitation capability of the algorithm.•It contributes to enhance the capacity of the algorithm in convergence and spread.