Abstract This paper concentrates on a resource-constrained multi-robotic disassembly line balancing (RC-MDLB) problem. In this RC-MDLB problem, different types of end-of-life products are disassembled simultaneously on the same line under the following conditions: allocating multiple robots to a workstation to simultaneously process the disassembly tasks that have no precedence relationship with each other, each robot needs a fixed number of limited resources to process tasks, and the total resources for each workstation is fixed. A mathematical model is presented for the RC-MDLB problem to minimize the cycle time and the number of robots being occupied simultaneously. A constrained multi-objective evolutionary algorithm framework and a constrained NSGA-II (E-NSGA-II) algorithm with epsilon method are proposed to handle the constraints of the RC-MDLB problem. The proposed E-NSGA-II is applied to a set of RC-MDLB problem instances introduced in this paper and compared with five representative multi-objective evolutionary algorithms. The experimental results reveal that the proposed E-NSGA-II presents outstanding performance on most of the cases analyzed.

中文翻译：

---

使用epsilon方法进行资源受限多机器人拆装线平衡的进化优化

摘要 本文重点研究资源受限的多机器人拆卸线平衡 (RC-MDLB) 问题。在这个 RC-MDLB 问题中，不同类型的报废产品在以下条件下在同一条线上同时拆卸：将多个机器人分配到一个工作站同时处理彼此之间没有优先关系的拆卸任务，每个机器人机器人需要固定数量的有限资源来处理任务，每个工作站的总资源是固定的。提出了 RC-MDLB 问题的数学模型，以最小化循环时间和同时占用的机器人数量。提出了约束多目标进化算法框架和带有epsilon方法的约束NSGA-II (E-NSGA-II)算法来处理RC-MDLB问题的约束。提出的 E-NSGA-II 应用于本文介绍的一组 RC-MDLB 问题实例，并与五种代表性的多目标进化算法进行了比较。实验结果表明，所提出的 E-NSGA-II 在分析的大多数情况下都表现出出色的性能。