基于连续时间MDP模型和随机决策的维护周期

Maintenance cycle based on continuous time MDP model and stochastic decision

生产系统的维护策略直接关系到设备的使用寿命,对生产线的连续性和可靠性,以及产品质量、生产效率、满足率等方面都有影响。利用连续时间的MDP模型研究单台设备的维护策略,综合考虑转移概率的动态性和方案选择的随机性,利用MATLAB软件实现优化,获得最佳维护周期。将其应用于混联结构,与基于离散空间的MDP维护策略进行比较。结果显示,基于连续时间MDP模型的维护策略能够有效提高生产系统性能,降低成本,提高生产效益。

The maintenance strategy of production system is directly related to life of equipment. And it effects the continuous and reliability of production line, product quality, production efficiency and satisfaction rate etc. This paper uses the continuous- time MDP model to study the maintenance strategy of single device. And it considers the dynamic of transition probability and the ran-dom of scheme selection, by simulation and optimization to obtain the optimal maintenance strategy in Matlab software. It is applied to the series and parallel structure, compared with discrete space MDP maintenance. The results show that maintenance strategy based on continuous- time MDP model can effectively improve performance of production system, reduce costs and improve produc-tion efficiency.