A novel bi-level multi-objective genetic algorithm for integrated assembly line balancing and part feeding problem. Issue 2 (17th January 2023)

• Record Type: Journal Article
• Authors: Chen, Junhao; Jia, Xiaoliang; He, Qixuan
• Is Part Of: International journal of production research. Volume 61:Issue 2(2023)
• Page Start: 580
• Page End: 603
• Publication Date: 2023-01-17
• Subjects: Integrated problem -- assembly line balancing -- part feeding -- bi-level multi-objective genetic algorithm -- Workload balancing -- Cost optimization; Factory management -- Periodicals; 658.57
• DOI: 10.1080/00207543.2021.2011464
• Languages: English
• ISSNs: 0020-7543

Abstract

Abstract : The manufacturing industry has been pursuing an efficient and economical assembly system. By considering assembly line balancing (ALB) and part feeding (PF) as an integrated problem and programming them simultaneously opens additional opportunities to improve the performance of the entire assembly system. However, the integrated ALB and PF problem is a non-deterministic polynomial (NP) hard problem. This implies that exact solutions cannot be obtained in a reasonable computation time and its near-optimal solutions can only be realised by meta-heuristics. In this study, we propose a novel bi-level multi-objective genetic algorithm (NBMGA) to solve the integrated ALB and PF problem. First, a bi-level mathematical model is established to simultaneously minimise the number of stations and workload smoothness of ALB in the upper level as well as the number of supermarkets of PF in the lower level. Second, the NBMGA with two modified strategies, including extending fitness evaluation and adaptive termination condition, is designed for problem solving. Finally, a series of computational experiments are conducted to demonstrate the efficacy of the proposed algorithm. The computational results indicate that the proposed algorithm outperforms the bi-level nondominated sorting genetic algorithm (NSGA) II in terms of the approximation to the true frontier without sacrificing computational efficiency.