Fault Diagnosis Strategy Optimization Under Unreliable and Multivalued Tests

In this paper, we propose a method of optimal diagnosis strategy based on Markov decision process with the inference process of fault diagnosis, and the multi-valued outcomes were treated as the independent signals. And, the uncertainty of tests is depicted by the fault detection and false alarm probabilities for the existence of multiple interference sources in real world. The simulation results also illustrate that this method is valid to get the optimal strategy with minimum cost.