A permutation-based algorithm to optimally reschedule trains in a railway traffic network

摘要

In this paper we discuss dynamic traffic management of railway traffic networks at an operational level. We design a model predictive controller based on measurements of the actual train positions. The core of the model predictive control approach is the railway traffic model, for which a switching max-plus linear system is proposed. If the model is affine in the controls, the optimisation problem can be recast as a mixed-integer linear programming problem. To this end we present a permutation-based algorithm to model the rescheduling of trains running on the same track. We apply the algorithm to a simple railway traffic network simulation model and show a significant reduction of delays compared to the uncontrolled case.