Modelling, performance evaluation and planning of public transport systems using generalized stochastic Petri nets

Owing to the tremendous size of public transportation systems, the design process of routes and timetables follows a stringent order. Centred on hub-based networks, this paper deals with two important phases of this process: the computation of the core line frequencies and departure times. To provide more reliable connections at a reasonable cost, we use generalized stochastic Petri nets. As a result, several improvements of the hub-based network scheduling are detailed. It is shown that the Petri net modelling is a powerful technique used to study the public transport systems performance (passenger waiting times) by comparing our results with deterministic ones. Finally, an application to a real size public transportation system is given to illustrate the results.