This paper addresses the problem of train conflict detection and resolution, which is dealt every day by traffic controllers to adapt the timetableto delays and other unpredictable events occurring in real-time. We describe a number of algorithmic improvements implemented in the real-time traffic management system ROMA (Railway traffic Optimization by Means of Alternative graphs), achieved by incorporating effective rescheduling algorithms and local rerouting strategies in a tabu search scheme. We alternate a fast heuristic and a truncated branch and bound algorithm for computing train schedules within a short computation time, and investigate the effectiveness of using different neighborhood structures for train rerouting. The computational experiments are based on practical size instances from a dispatching area of the Dutch railway network and include complex disturbances with multiple late trains and blocked tracks. Several small instances are solved to optimality in order to compare the heuristic solutions with the optimum. For small instances, the new tabu search algorithms find optimal solutions. For large instances, the solutions generated by the new algorithms after 20 s of computation are up to more than 15% better than those achieved within 180 s by the previous version of ROMA. (A) Reprinted with permission from Elsevier.
Samenvatting