On-demand ridesharing with optimized pick-up and drop-off walking locations

On-demand systems in which passengers with similar routes can share a vehicle are expected to become relevant part of future mobility, thanks their flexibility and potential impact on reducing congestion. Nevertheless, due the long detours required by door-to-door scheme, they induce extra costs users terms delay. In this paper, we face design such system might be requested online walk towards/from nearby pick-up/drop-off points if improves overall efficiency. We show theoretically that general problem becomes more complex (as it contains two sub-problems extend set-cover), analyze trade-offs emerge, provide formulation specific heuristics able solve over large instances. test real dataset Manhattan taxi trips (9970 requests during one hour), finding (a) average walks about minute reduce number rejections than 80% Vehicles-Hour-Traveled 10%, (b) who depart or arrive at most demanded areas likely walk, (c) performance improvement service is larger when receives trip requests.