Optimisation of train re-scheduling using a genetic algorithm applied to the mimic panel state model of the overall rail network

The re-scheduling of trains in response to a disruptive incident is an important, yet complicated task, and there is a very strong need for the development of on-line decision support systems for the task. However, the optimisation of train re-scheduling, which is necessary for devising such a system, is computationally very hard. It is therefore important to develop a model on which optimisation can be performed with enough computational efficiency. In this paper, the authors present a new and simplified model for the purpose. It is named the "mimic panel state model", because it is intended as the mathematical representation of a "state" shown in the mimic panel that can be found in many actual train control rooms. The model is described in detail, and an example application of this model to the numerical solution of train control (or train re-scheduling) problems is presented.