Optimal Multi-Type Inspection Policy For Systems With Imperfect Online Monitoring

Both online monitoring and manual inspection are widely used in identifying a system's health state, based on which proper preventive maintenance (PM) can be carried out. The existing maintenance optimization models typically consider only online monitoring or a single type of inspection and assume they can perfectly reveal the system's state. However, the information provided by online monitoring is never perfect, and it needs to be combined with the inspection to identify the system's health state. Besides, there are usually many types of inspections with different costs and detecting capabilities in real applications. It is hence equally important to decide on when operators should make an inspection and which kind of inspection should be performed based on the information from online monitoring. In this paper, we formulate the multi-type inspection decision-making problem within the partially observable Markov decision process (POMDP) framework. We establish some structural properties for the optimal multi-type inspection policy. The value iteration and the lambda-minimization algorithm are combined to obtain the optimal solution. A case study is provided to illustrate the optimal inspection policy and its advantage over some existing policies.