Design and Implementation of Algorithm Based Railway Locomotive and Rolling Stock Maintenance Management System

In view of the problems of low maintenance decision-making efficiency, unreasonable resource allocation and inaccurate fault prediction in the current maintenance management of railway locomotives and vehicles, this paper references and implements a railway locomotive and vehicle maintenance management system based on deep neural network (DNN). First, the DNN model is used to comprehensively analyze and evaluate the operating status of locomotives and vehicles. Then, the powerful learning ability of DNN is used to identify and predict fault modes, so as to warn potential faults in advance. Finally, the optimization algorithm is combined to dynamically schedule and allocate maintenance resources to ensure efficient utilization of resources and timely implementation of maintenance work. The research results show that the accuracy of the DNN model in fault prediction is significantly better than other algorithms, and its AUC (Area Under the Curve) value reaches 90%. In addition, the resource scheduling method optimized by genetic algorithm reduces the task completion time by 10%, increases resource utilization by 10%, and significantly reduces maintenance costs. In the above data conclusions, the research system provides an effective solution for the intelligent maintenance management of railway locomotives and vehicles.