A green and efficient disassembly line balancing with human-robot collaboration and destructive disassembly

Abstract

Human-robot collaboration combines the strengths of both humans and robots to enhance disassembly line efficiency. Considering the indivisibility and recovery value of certain components, this study incorporates destructive disassembly into a human-robot collaborative disassembly line. A mixed-integer linear programming model of the disassembly line balancing problem is constructed. The model accounts for task precedence relationships, task attributes, disassembly modes, human-robot collaboration, and the configuration of humans and robots. The objective is to minimize cycle time, smoothness index, and disassembly energy consumption while maximizing disassembly profit. The algorithm uses a three-layer encoding strategy based on task sequence, task operators, and disassembly modes, with an optimization-driven initialization to improve the initial solution quality. Five selection strategies and two neighborhood search strategies are designed, and during the iterative process, the strategy is dynamically adjusted through Q-learning to enhance both global search and local search capabilities. The effectiveness and superiority of the proposed algorithm are validated through three types of test case experiments, compared with the five latest algorithms. Finally, the model and algorithm are applied to a real-world laptop disassembly case. The results show that the introduction of collaborative robots in disassembly significantly reduces disassembly costs. Compared to manual disassembly, the cycle time of the disassembly line can be reduced by 24.39%, and idle time can be reduced by 33.64% in the human-robot collaborative disassembly mode. Compared to non-destructive disassembly, destructive disassembly can reduce energy consumption by 28.20%.