Improved optimisation method considering full solution space for disassembly line balancing problem in remanufacturing system

Abstract

Disassembly is a key process towards waste recycling. The disassembly line balancing problem (DLBP) is a hot topic in achieving efficient disassembly. However, the existing decoding approaches cannot provide the best disassembly plans because solution spaces were reduced when solving the DLBP. To provide efficient recycling schemes for decision-makers, this study designed stochastic and partial stochastic decoding approaches searching the full solution space for four classical layouts: straight DLBP (SDLBP), U-shaped DLBP (UDLBP), two-sided DLBP (TDLBP), and parallel DLBP (PDLBP). In addition, this study constructed four mixed-integer programming (MIP) models considering both adjacent task constraints and tool switching constraints between adjacent tasks and between the first and final tasks of adjacent cycle times, optimising the number of workstations, idle time balancing index, number of tool switches, and energy consumption for the SDLBP, UDLBP, TDLBP, and PDLBP. Then, the MIP models as well as the stochastic, partial stochastic, and original decoding approaches were applied to solve four disassembly layouts. Finally, the comparison results show that the designed decoding approaches have superior solution performances and can recycle waste products more efficiently than existing methods. In addition, three heuristic algorithms, the NSGA-Ⅱ, the artificial fish swarm algorithm, and the differential evolution algorithm, combined with the designed partial stochastic decoding approach and have excellent solving performance, were applied to optimise a waste printer industrial case. Then, the optimisation results in obtaining balanced operation schemes for each objective.