中文摘要：

目前，复杂系统的修复和检测费用往往比较高，过程比较繁琐，且由于设备故障停机造成了大量损失。针对这一问题，提出一种通过系统产出间接地判断整个系统是否存在问题，并在此基础上进行预防性维修的决策方法。与传统检测的方法不同，利用控制图监控产品，充当设备的传感器，间接地判断生产产品的设备是否正常。根据不完美检测下的监控结果，进行适当的维修策略，基于更换过程理论，建立单位时间成本最小的目标函数，提出控制图与预防性维修的优化参数设置。通过遗传算法进行数值仿真，表明该模型能有效实现控制图与预防性维修的联合经济设计目标。研究对于生产系统节约成本、提高综合经济效益具有现实意义。

英文摘要：

Currently the repair and maintenance of complex systems are usually complicated and costly, and the breaking-down equipment would cause great losses. Focusing on this situation, a decision-making method that can judge the system indirectly by the system output was proposed. Being different from traditional detection methods, control chart of monitor products acts as a sensor device and indirectly judges whether the equipment of the production is normal. The optimal preventive maintenance policy was obtained by minimizing cost per unit time through the given control charts of the imperfect detections. Finally, a numerical example based on genetic algorithm was given to illustrate the calculation process and the availability of the proposed PM model. It also shows that the product system cost under the proposed PM strategy is lower than that traditional detection strategy.