Stochastic and robust optimization of on-demand transport systems linked to public transport

Traditional public transport systems provide efficient mass transportation solutions between fixed service points. By definition, these systems can only serve a fixed area. Consequently, passengers who request transportation services outside this region have reduced access to public transportation systems. On-demand transportation systems overcome this problem by offering shuttle services to serve passengers that face such restrictions. This study focuses on building optimization models to obtain cost-optimal solutions for on-demand transportation systems integrated with railway systems, where the fixed service point is defined as a specific train station. We develop robust solutions by ensuring a predefined level of feasibility over a scenario set that represents travel times. This can be done by implementing chance constrained programming to integrate scenario-based travel times into the model, which in return, builds robust and optimal solutions for a specific percentage of scenarios.