Data-Driven Optimization of Railway Track Inspection and Maintenance Using Markov Decision Process
MetadataShow full item record
The goal of this thesis is to develop a data-driven condition-based maintenance policy for the track inspection. This thesis will help in maintaining high service level of the railway tracks which is a difficult task to accomplish. We use a two-year track geometry inspection dataset which contains a variety of geometry measurements for every foot. We separate the data based on the time interval of inspection run, calculate the aggregate TQI for each section, develop a Markov Chain for modeling track deterioration, build a Markov Decision Process (MDP) for track maintenance decision making and optimize it using value iteration algorithm. By comparing with existing maintenance policy with Markov Chain Monte Carlo (MCMC) simulation, the new maintenance policy developed in this thesis results in saving around 10% maintenance costs.