A Multi-objective Two-sided Disassembly Line Balancing Optimization Based on Artificial Bee Colony Algorithm: A Case Study of an Automotive Engine

Disassembly is an important part of green manufacturing and remanufacturing. The disassembly line is an optimum form for mass and automatic disassembly in the industry. To optimize the disassembly system and the use of resources, the disassembly line balancing problem (DLBP) has attracted much attention. Compared with the conventional one-sided straight disassembly line, the two-sided disassembly line can use both the left and right sides of a conveyor belt for disassembly operation, thereby improving production efficiency, especially for large-sized and complicated products. On the other hand, due to the constraints and precedence among parts, it is a challenge to plan the disassembly scheme for a two-sided disassembly line. In this paper, a model is established to solve a two-sided disassembly line balancing problem (TDLBP). First, a hybrid graph is utilized to express constraints and precedence relationships, and a novel encoding and decoding method is developed for the disassembly scheme planning of a two-sided line for handling the challenge caused by constraints and precedence among parts. Then, a multi-objective TDLBP optimization model is proposed including the number of mated-workstations, idle time, smoothness index, the auxiliary indicator, and a meta-heuristic based on an artificial bee colony (ABC) algorithm is designed to solve TDLBP. Finally, the proposed model and method are applied to an automotive engine case, and the results are compared with NSGA-II, hybrid artificial bee colony algorithm (HABC), and multi-objective flower pollination algorithm (MOFPA) to verify the practicality of the proposed model in solving the TDLBP.