In enterprises, machines are subject to some unavailable periods resulting in production halts. The consideration of preventive maintenance (PM) and resource assignment in straight and U-shaped assembly lines is of paramount importance. Hence, this paper aims to tackle the integration of PM and balancing problems in straight and U-shaped resource-dependent assembly lines via two new mixed-integer linear programming (MILP) models and a memetic algorithm. In the MILP models, the optimal PM intervals of all machines and all PM scenarios are determined according to the maximization of the availability. In the memetic algorithm, a multi-attribute solution representation, including heuristic rule matrix, task separator set and machine assignment permutation, is designed to build a feasible solution. For this multi-attribute solution representation, this algorithm respectively designs different crossover, mutation and local search operators to exploit the space around the solutions. Finally, five comparison experiments are conducted to prove the effectiveness of the MILP models in small-scale instances and the superiority of the improvements and memetic algorithm in all instances.

在企业中，机器会受到一些不可用时段的影响，从而导致生产停顿。在直线和 U 形装配线中考虑预防性维护 (PM) 和资源分配至关重要。因此，本文旨在通过两种新的混合整数线性规划 (MILP) 模型和模因算法来解决直线和 U 形资源依赖型装配线中 PM 的集成和平衡问题。在 MILP 模型中，所有机器和所有 PM 场景的最佳 PM 间隔是根据可用性的最大化来确定的。在模因算法中，设计了一个多属性的解决方案表示，包括启发式规则矩阵、任务分隔符集和机器分配排列，以构建可行的解决方案。对于这个多属性解表示，该算法分别设计了不同的交叉、变异和局部搜索算子，以利用解周围的空间。最后，通过五个对比实验来证明 MILP 模型在小规模实例中的有效性以及改进和模因算法在所有实例中的优越性。