An efficient greedy heuristic for the real-time train platforming problem

Abstract

This paper focuses on a special class of real-time railway traffic management problem: efficiently reallocating trains to platforms at stations in case of slight schedule changes or major disruptions. The need for a real-time response with a high level of quality makes this problem particularly challenging.

To address this issue, the authors propose a mesoscopic approach that involves preprocessing the data to determine feasible routes and other disruption parameters. They develop a greedy interchange heuristic to solve the mesoscopic real-time train platforming problem, providing high-quality routing and timing decisions within the computational time constraints of real-time management problems.

The performance of the proposed heuristic is analyzed through case studies using both synthetic and realistic scenarios from the Spanish railway traffic system. For large instances of the Atocha-Cercanías station case study, the solutions are generated from 5 to 10 times faster by the heuristic algorithm than by the exact method. The authors conclude that the proposed heuristic is a promising solution for real-time train platforming problems.