中文摘要：

为了减少串行系统中设备停机造成的经济损失,针对非马尔科夫串行生产系统提出一种预防性维护与缓冲区分配联合优化模型.根据串行系统的运作机制,构建了系统在制品的加工开始一释放时间递推式及对应产出率约束,推导了系统运营中的维护成本、故障小修成本与在制品持有成本,建立了基于设备故障抽样的仿真模型评价系统性能.以最小化系统运营的各项成本为优化目标,建立了有限时间计划的联合优化模型.设计了基于惩罚函数的遗传算法来求解最优预防性维护计划与缓冲区分配,通过实例分析表明了联合优化模型及其算法的有效性.

英文摘要：

To reduce the economic losses caused by breakdown of machines, a model for joint optimization of preven- tive maintenance and buffer allocation was proposed for non-Markovian serial system. The recursion formula of pro- duction start and release time, and the constraint of throughput rate were set up according to a serial production system＇s mechanism, from which the costs of maintenance, repair and inventory were derived. A failure sampling simulation model was established to evaluate the system＇s performance. Based on the descriptions mentioned above, a joint optimization model was established with an objective of minimizing the expected total cost in a finite time ho- rizon. A genetic algorithm with penalty function to solve the optimal preventive maintenance plan and buffer capaci- ties was developed. The results indicated that the joint optimization model was valid and efficient.