Stochastic optimization of public transport schedules to reduce transfer waiting times

We address the problem of transit schedule optimization in a multi-modal transport system. Using a real-world collection of transit instances in a transport system, we revise the two-stage stochastic program for transit schedule optimization initially proposed in [10] to improve the theoretical schedule and reduce the passengers' expected waiting times. We address some limitations of the approach when applying to real transit events, and propose a more general solution which gets rid of some restrictive assumptions, for the price of a small complexity increase. We report results of testing the optimization algorithm on the transit data collected in the public transport system of Nancy, France.