A closed queuing network based modeling and joint optimization of the maintenance system using a modified genetic algorithm

Abstract

This research develops a joint optimization model for the maintenance system of railway locomotives based on a closed queuing network, so as to optimize the maintenance resources allocation and improve the system performance in operation, maintenance and spares stages. The closed queuing network is applied to determine the locomotives availability level, and then is integrated into a cost optimization model. By applying a modified genetic algorithm, the optimal (or approximate) quantities of maintenance channels and operating locomotives are achieved upon high system availability at minimal costs. Based on the data collected from railway application, it is shown that the proposed joint optimization model leads to a significant improvement in both resources allocation and the system performance. Taguchi design method is used afterwards to analyze the independent and interaction effects of relevant parameters on the entire maintenance system.