Balancing multi-product disassembly lines under uncertainty: an enhanced multiobjective bees algorithm

Disassembly is a crucial initial step in the remanufacturing and recycling of products that reach the end of their service life. The growing emphasis on green remanufacturing and recycling underscores the importance of efficient disassembly lines. Multi-product disassembly lines can accommodate a wider range of products, enhancing resource utilisation and overall efficiency. Consequently, the multi-product disassembly line balancing (DLB) problem has garnered increasing attention. However, current research on the multi-product DLB problem largely focuses on deterministic information, which fails to reflect the complexities of real-world operations. Moreover, existing approaches to uncertain DLB problems have limitations. They often rely on triangular fuzzy sets and stochastic simulation methods, which may oversimplify uncertainty or depend on known complete probability distributions. To better align with the disassembly environment in the real world, this study investigates the multi-product DLB problem under uncertain conditions, using trapezoidal fuzzy numbers and developing a multiobjective model. Our model aims to minimise the idle rate of the disassembly line, reduce the hazard index, lower energy consumption, and optimise the number of workstations. To tackle this problem effectively, we introduce an enhanced multiobjective bees algorithm (EMOBA), which combines innovative bee foraging strategies, an adaptive operator selection method based on reinforcement learning, and a local search strategy. Experimental results demonstrate that EMOBA significantly outperforms existing algorithms. To the best of our knowledge, this research marks the first exploration of this specific problem, laying the groundwork for future studies in this field.