An incremental decision algorithm for railway traffic optimisation in a complex station

In heavy traffic nodes of rail networks, as in main stations, conflicts and subsequent train delays greatly propagate during operations. A disturbance which is at origin only a few seconds long can quickly lead to other delays of over five minutes. To limit this phenomenon, newmethods andmodels are necessary to optimise the use of scarce resources like platforms and track sections. The operator controlling traffic has to select and evaluate alternative solutions that reduce delays caused by conflicts. This task can be formulated as an optimisation problem, where the decision variables are the selection of the alternative routes and sequences for trains and the criterion is the sum of delays. This optimisation problem is a joint scheduling and allocation problems and is considered to be a NP-hard problem which makes it hard to solve using exact methods for a reasonable problem size. In this paper, we compare two heuristic methods for solving the problem. The first one uses a two-phase approach to perform independently resource allocation and scheduling. The second one performs incrementally the two kinds of decisions at track section level, i.e. at each step, the algorithm performs decisions of allocation of a track section or scheduling a pair of train runs on this section. The algorithm uses a measure of the criticality of the track sections to heuristically choose between making resource allocation decisions or scheduling decisions.