The dial-a-ride problem (DARP) aims at providing solutions to on-demand col- lective human transportation problems. It generalizes the pickup and delivery problem with time windows (PDPTW) by considering customer convenience constraints in addition to standard PDPTW constraints. The DARP with Transfers (DARP-T) has recently been in- troduced by Péton et al : [10] in order to permit the customers to change vehicles during their trip, at predefined transfer points. One practical application of the DARP-T is an on-demand public transportation service planner, provided a solution framework dealing with dynamic instances. In this paper, we explore several Stochastic Programming models to the dynamic problem. By modeling customer demands as stochastic data, we can compute anticipative solutions which are dynamically adapted online in order to maximize the number of online satisfied requests. We also propose a heuristic algorithm that can be exploited on any Vehicle Routing problem, including the Dial-a-Ride Problem with Transfers. This paper does not provide any results, but aims at describing a research plan for the current PhD thesis.