Applying Markov Decision Processes to Maintenance with Product Selection

Abstract

There is a lot of previous research on maintenance problems. In some previous research, an expected profit is maximized considering sales amount of product. In the previous research, product selection is not considered. The expected profit depends on product selection when sales amount and equipment state transition probabilities for operation (probabilities of equipment deterioration) of each product are different. In this research, maintenance problem with product selection is studied. Sales amount and equipment state transition probabilities for operation of each product are different. The maintenance problem with product selection is modeled by Markov decision processes. A new maintenance method which maximizes the expected profit based on statistical decision theory is proposed. In the proposed method, dynamic programming method is used. The effectiveness of the proposed method is shown by some computational examples. The expected profit of the proposed method is greater than that of a comparison target. In this research, the expected profit is maximized under the condition that all probabilities are known. But the probabilities are unknown in real cases. An expansion of this research with unknown probabilities is one of further works.