Scheduling of energy-efficient distributed blocking flowshop using pareto-based estimation of distribution algorithm
2022; Elsevier BV; Volume: 200; Linguagem: Inglês
10.1016/j.eswa.2022.116910
ISSN1873-6793
AutoresXiaohui Zhang, Xinhua Liu, A. Cichoń, Grzegorz Królczyk, Zhixiong Li,
Tópico(s)Assembly Line Balancing Optimization
ResumoThis study investigates the impact of production scheduling decisions aims at improving productive and energy-efficient performances simultaneously in distributed blocking flowshops (EDBFSP). To reach a compromise between the conflicting objectives, a Pareto multi-objective optimization model based on the estimation of distribution algorithm (MOEDA) is proposed. Firstly, an initialization method based on the problem-specific characteristics is designed to create a promising population with quality and diversity; secondly, a probabilistic model based on a Bayesian network is constructed to predict position relationships between jobs. Two neighborhood operators with modified insertion technique are proposed to realize the adjustments of both job sequence and processing speed; thirdly, two operators are developed to execute multi-objective local searches on the elite solutions. Aiming at efficient utilization of the resulted blocking and idle time, an energy-saving method is designed for EDBFSP. In the experimental parts, to gain the best performance, the key parameters of MOEDA have been calibrated. The validation is conducted to assess the performances of the designed initialization method, neighborhood search, local search, and energy-saving strategies. The proposed MOEDA is also compared with mainstream metaheuristics for solving green scheduling problems. The experiment results show that the optimization and search ability of MOEDA have gained prominent advantages over other metaheuristics in both precision and distributivity.
Referência(s)