REPLACEMENT POLICY FOR A PARTIALLY OBSERVABLE MARKOV DECISION-PROCESS MODEL USING FUZZY DATA

When the deterioration is governed by a Markov process, such processes are known Partially Observable Markov Decision Processes (POMDP) which eliminate the assumption that the state or level of deterioration of the system is known exactly. This research investigates a two state partially observable Markov chain in which only deterioration can occur and for which only actions possible are to replace or to leave alone. Because the cost structure is defined on the process can not be obtained exactly, a fuzzy concept is proposed for an optimal replacement policy which leads to the operation cost savings of a machine. The goat of this research is to develop optimal replacement policies under a new approach which has the potential for solving other problems dealing with continuous state space Markov chains using fuzzy data.