Optimum maintenance policy with Markov processes

This paper presents a method to find the optimum maintenance policy for a component. Random failures and failures due to deterioration are considered. Using Markov processes, the state probabilities are calculated and the optimal value of the mean time to preventive maintenance is determined by maximizing the availability of single component with respect to mean time to minimal preventive maintenance. Using the state probabilities, the problem is set up as Markov decision processes and an optimum maintenance policy using the policy iteration algorithm is determined. An example is used to illustrate the method. Maple V and Matlab software have been used to solve the equations. (c) 2005 Elsevier B.V. All rights reserved.