A MARKOVIAN MODEL AS A TOOL FOR OPTIMISATION OF MAINTENANCE INTERVENTIONS ON RAILWAYS LINES

This paper is designed to develop a procedure which defines a planning criterion for railway superstructure maintenance by means of Markov decision processes. This methodology allows to formulate a specific policy  which carries out the best configuration of budget allocation (min. Ф), and at the same time to guarantee the highest efficiency level in the railway superstructure. Thanks to the dynamic programming technique applied to decision processes, the report has examined the possibility of establishing the best management policy in order to maintain adequate safety levels of implementation and quality speed levels, in the presence of budget constraints, thus optimizing the available resources. This procedure therefore allows to optimize the maintenance protocols currently adopted which ensure the superstructure efficiency state through an "on condition" maintenance, based on the surveys of the track geometric parameters conducted on high-performance diagnostic carriages. Finally, a case study has been developed in order to estabilish the application limits of the methodology suggested in order to formulate the best strategy for managing the resources and improving speed quality and safety.