Disassembly sequence planning of end-of-life smartphones based on an improved PSO-GA algorithm

Efficient and damage-free disassembly of components is one of the effective methods for achieving resource reutilization and sustainable development. For disassembly sequence planning of end-of-life (EoL) smartphone parts, intelligent algorithms with rapid convergence and dynamic optimization capabilities are leveraged to enhance disassembly efficiency. However, the Particle Swarm Optimization-Genetic Algorithm (PSO-GA) exhibits limitations in dealing with problems with complex constraints. This drawback mainly arises from excessive redundant search paths generated during iterations, which drastically increases iterative overhead and deteriorates the convergence rate of the traditional PSO-GA. To overcome this challenge, an improved PSO-GA integrating the gravitational search mechanism is proposed to optimize the disassembly sequence planning. This algorithm is based on a genetic crossover-mutation combined optimization mechanism, which overcomes the limitations of algorithmic optimization and effectively expands the search range for optimization. The algorithm replaces the particle vector position constraints in traditional PSO by constructing a model of the structural connection matrix and the process constraint matrix. The improved PSO-GA algorithm is validated using the iPhone 8 as a case study. The case studies demonstrate the reliability of the improved PSO-GA algorithm in the following two aspects: its convergence speed in the optimization of disassembly energy consumption and profit. In particular, compared with the traditional PSO-GA, the convergence speed of the improved PSO-GA is increased by 61% and 65%, respectively. This work verifies the improved algorithm’s effectiveness and applicability in disassembly sequence planning.