Predicting the matching probability and the expected ride/shared distance for each dynamic ridepooling order: A mathematical modeling approach

• A mathematical modelling approach is proposed to predict the matching probability and expected ride/shared distance of each ridepooling order. The model explains complex mechanism that determines OD pair. generates fairly good predictions under diverse conditions demand intensities. popularity smartphones advent GPS positioning wireless communication technologies in recent decade have facilitated large-scale implementations dynamic services, such as Uber Pool, Lyft Line, Didi Pinche. As services trips usually start before appearance pooling partners, knowing getting with another order (i.e., probability), detour distance, shared trip essential for passengers evaluate their willingness pool platforms offer attractive discounts. In this paper, assuming every passenger shares vehicle space at most one during entire trip, orders (origin-destination) pair appear following a Poisson process given rate, we propose modeling probability, ride first-come-first-serve strategy service. method defines unmatched different locations along exclusive-riding path into seeker- taker-states, formulates interdependency probabilities, rates arrival (unmatched) states system nonlinear equations, probabilities distances all pairs simultaneously. Under same strategy, simulated occurrence, movements state transitions based on 30*30 grid network real Haikou City China. comparison simulation results, show paper can generate satisfactory