Abstract: Conventional unrelated parallel machine scheduling typically ignores machines' reliability loss during machine processing,resulting in production interruptions due to excessive machine deterioration.Aiming at this problem,an integrated scheduling model that joint the scheduling scheme and Preventive Maintenance（PM）strategy was constructed by taking the jobs allocation and the machine working age threshold as the decision variables,and a hybrid discrete particle swarm algorithm（HDPSO）was designed to solve it.HDPSO redefined the particle position update mechanism in the solution space based on two-dimensional coding,and a multi-neighborhood search strategy was incorporated to balance the exploration and development capabilities in the discrete solution space,which achieved the synergistic optimization of scheduling and PM in the maximum completion time index.Extensive simulation demonstrated that HDPSO had a high effectiveness in solving large scale unrelated parallel machine scheduling problems.Meanwhile,the proposed working age maintenance strategy was more conducive to increase productivity and scheduling performance than periodical maintenance.

Key words: preventive maintenance, unrelated parallel machine, working-age maintenance, discrete particle swarm optimization algorithm

摘要： 传统不相关并行机调度通常忽略机器可靠度在加工过程中的损耗,导致机器因劣化过度而带来生产中断。针对该问题,以工件分配与机器役龄阈值为决策变量,构建将调度方案与维护策略进行协同优化的集成调度模型,并设计一种混合离散粒子群算法对模型进行求解。该算法基于二维编码方式对粒子的位置更新机制进行重新定义,并融入多邻域搜索策略平衡算法在离散解空间内的探索与开发能力,实现调度与维护二者在最大完工时间指标上的协同优化。大量仿真结果验证了所提算法在求解大规模不相关并行机调度问题方面具有较强的搜索优势。同时,所提的役龄维护策略相较于定周期维护而言更有助于提升生产效率,改善调度性能。

关键词: 预防性维护, 不相关并行机, 役龄维护, 离散粒子群算法