To secure a larger market share in the dynamically evolving personalized market demand, enterprises have to adopt a more flexible production mode. In this context, the implementation of parallel production with multiple product types necessitates frequent switching between different products, imposing elevated requirements on machine fault-free operation and stable production. There is a compelling motivation to investigate the joint optimization of the flexible job shop scheduling problem and preventive maintenance (FJSSP-PM). However, existing research has primarily focused on utilizing predetermined maintenance as inputs for optimizing production scheduling, while overlooking the uncertain maintenance requirements of machines during the production process. Therefore, this paper proposes a two-stage joint optimization model to simultaneously address three subproblems: machine assignment, operation sequencing, and maintenance arrangement. Specifically, the first stage involves constructing a mathematical model for the FJSSP to minimize penalties for tardiness and workload balancing. Subsequently, a comprehensive failure rate model is developed to determine maintenance requirements for machines with production-dependent failure behavior, particularly under high-frequency production switching scenarios. Furthermore, a feedback-updated strategy is proposed to achieve synchronous optimization of production and maintenance activities in an economically and operationally efficient manner. Given the complexity of the problem, a hybrid meta-heuristic algorithm is designed to effectively handle large-scale cases. The effectiveness of the proposed model in terms of cost-savings has been verified compared with existing relevant methods.

为了在动态变化的个性化市场需求中获得更大的市场份额，企业必须采用更加灵活的生产模式。在此背景下，实现多品种并行生产需要频繁切换不同产品，对机器无故障运行和稳定生产提出了更高的要求。研究灵活作业车间调度问题和预防性维护（FJSSP-PM）的联合优化有令人信服的动机。然而，现有研究主要集中于利用预定维护作为优化生产调度的输入，而忽视了生产过程中机器的不确定维护需求。因此，本文提出了一种两阶段联合优化模型，以同时解决三个子问题：机器分配、操作排序和维护安排。具体来说，第一阶段涉及为 FJSSP 构建数学模型，以最大限度地减少迟到和工作负载平衡的惩罚。随后，开发了一个全面的故障率模型，以确定具有与生产相关的故障行为的机器的维护要求，特别是在高频生产切换场景下。此外，还提出了反馈更新策略，以经济且高效的方式实现生产和维护活动的同步优化。考虑到问题的复杂性，设计了一种混合元启发式算法来有效处理大规模情况。与现有相关方法相比，验证了该模型在节省成本方面的有效性。