Imperfect preventive maintenance optimization for flexible flowshop manufacturing cells considering sequence-dependent group scheduling. (August 2018)

• Record Type: Journal Article
• Authors: Feng, Hanxin; Xi, Lifeng; Xiao, Lei; Xia, Tangbin; Pan, Ershun
• Is Part Of: Reliability engineering & system safety. Volume 176(2018)
• Page Start: 218
• Page End: 229
• Publication Date: 2018-08
• Subjects: Genetic algorithm -- Group scheduling -- Machining condition -- Preventive maintenance -- Simulated annealing; Reliability (Engineering) -- Periodicals; System safety -- Periodicals; Industrial safety -- Periodicals; Fiabilité -- Périodiques; Sécurité des systèmes -- Périodiques; Sécurité du travail -- Périodiques; 620.00452
• DOI: 10.1016/j.ress.2018.04.004
• Languages: English
• ISSNs: 0951-8320

Abstract

Highlights: Proposing a modified machine's reliability model under group-varying conditions. Proposing a cost-based model for integrating sequence-dependent GS and imperfect PM. Providing a simulated annealing embedded genetic algorithm to solve the problem. Our PM policy saves cost when compared with three other PM policies. The proposed approach outperforms genetic algorithm and simulated annealing. Abstract: Flexible flowshop manufacturing cells (FFMCs) have been widely applied to achieve the efficiency of high-volume manufacturing for products of small-to-medium demand. In FFMCs, jobs from several groups are processed at several stages with the same route and group-varying machining conditions. This paper integrates imperfect preventive maintenance (PM) and sequence-dependent group scheduling (GS) in FFMCs. A machine-level model is developed to describe machine reliability evolution under group-varying conditions. A system-level model is proposed to simultaneously obtain the planning of PM and GS by minimizing PM cost, minimal repair cost, and job tardiness cost in FFMCs. To solve the model, a simulated annealing embedded genetic algorithm (SAGA) is developed. Experiments show that the PM policy obtained by our proposed model can remarkably save cost when compared with three other types of policies. Further experiments present that the integer-key based chromosome representation of SAGA excels a random-key based representation, and SAGA outperforms SA and GA in terms of both the solution quality and robustness.