Applying the Markov decision processes to model changes to the maintenance states of an object

Operation and maintenance of technical objects is related to occurrence of various events, the effects of which affect the operation and maintenance process course, and particularly the courses of their use and service processes. Occurrence of those events may be of both determined and random nature. Assessing, analysing and forecasting the operation and maintenance process course, in complex technical systems, are connected with the problems of modelling the operation and maintenance processes of technical objects. Those processes are random ones which depend on one another. The paper presents an example of the Markov decision process to model changes of the analysed operation and maintenance states of technical objects. The investigation object is a real operation and maintenance system of urban transport buses in a middle-sized agglomeration (about 400 k residents). Supporting a decision maker, in the process of making decisions concerning performance of the operation and maintenance process of the means of transport, may be carried out by analysing the results of the investigations of the operation and maintenance process model. The investigations of that type are to determine values of the selected measures of technical and economic efficiency of the process being carried out for the estimated values of the model parameters. The values of the model parameters were estimated on the basis of the analysis of the results of the investigations performed in the analysed system of urban bus transport. A change to the values of the model parameters may reflect a change of influence of internal and external factors on behaviour of the system and the operation and maintenance process of the means of transport being carried out in it. Mathematical models of the operation and maintenance processes are intrinsically simplified, therefore practical conclusions resulting from investigating those models should be formulated carefully.