针对多状态半马尔可夫退化系统,研究了系统发生故障后的维修模型.在系统的退化过程中,有多个不可逆的工作状态和故障状态.系统的所有故障均能修复且可忽略故障修复时间,修复后系统回到正常的工作状态.当系统发生故障时,系统的故障检测设备首先确定出当前所处的故障状态,然后采取相应的维修措施.基于系统长期平均成本率最低的决策准则,运用半马尔可夫决策过程的策略迭代算法给出了系统处于不同故障状态下的最优维修策略.以实例说明了求解最优故障维修策略的迭代过程.
A failure maintenance model is studied for multi-state semi-Markovian deteriorateing systems. During deterioration process, a system has a finite number of inverse working states and failure states. All failure states of system can be repairable and the failure repair time is negligible. The system is in a normal working state after restore. Its failure-examination devices can detect the failure state immediately as soon as the failure occurs, and then the corresponding maintenance actions are carried out. The optimal failure-repair policy that minimizes the expected long-run cost rate of the system is obtained by using policy-iteration algorithm of semi-Markovian decision process. Finally, numerical examples are given to illustrate the iteration process for solving the optimal failure maintenance policy.