Assembly line balancing is an important and well recognised operations research problem. The current line balance may not stay optimal, even feasible, due to the disruptions in one or more workstations. In this study, after the disruption, we aim to rebalance the assembly line by considering the trade-off between workload balancing (fairness measure) and total replacement distance for the tasks assigned to the different workstations (stability measure).

We try to generate all non-dominated objective function vectors for the defined fairness and stability measures. Two algorithms are developed: exact algorithm (classical approach) and tabu search algorithm. The results of the experiments have shown that the classical approach returns exact non-dominated objective vectors with up to 40 tasks and 7 workstations in one hour, and the tabu search algorithm returns approximate non-dominated objective vectors that are very close to their exact counterparts and can solve large sized instances with up to 94 tasks and 7 workstations in less than 10 s.

装配线平衡是一个重要且公认的运筹学问题。由于一个或多个工作站的中断，当前的生产线平衡可能无法保持最佳，甚至不可行。在这项研究中，在中断之后，我们的目标是通过考虑工作负载平衡（公平性度量）和分配给不同工作站的任务的总替换距离（稳定性度量）之间的权衡来重新平衡装配线。

我们尝试为定义的公平性和稳定性度量生成所有非支配目标函数向量。开发了两种算法：精确算法（经典方法）和禁忌搜索算法。实验结果表明，经典方法在一小时内返回精确的非支配目标向量，最多 40 个任务和 7 个工作站，禁忌搜索算法返回非常接近其精确对应物的近似非支配目标向量并且可以在不到 10 秒的时间内解决多达 94 个任务和 7 个工作站的大型实例。