Markov Processes with Fuzzy Parameters - A Case Study

Detailed maintenance modeling is indispensable for utilities to determine optimum maintenance policy. Traditional reliability studies assume that transition rates or probabilities in Markov models are accurate. However, in reality, reliability data is either insufficient or mixed with uncertainty. This paper intends to utilize fuzzy set theory to represent parameters for Markov and semi-Markov processes. Previous single equipment maintenance models are extended with fuzzy transition parameters in Markov processes. The sensitivity analysis is performed to determine how fuzzy membership functions and boundary ranges impact equipment availability. Results are also compared with tradition non-fuzzy method. This work is valuable for utilities to develop maintenance models with incomplete and uncertain reliability data.