Fixed-route transit in the U.S.A. is facing stiff competition from more convenient mobility options. Microtransit—shared transportation that offers dynamic routing and scheduling to efficiently match demand—is emerging as an ally to fixed-route services. However, its positive impacts are too often constrained by the politics and economics imposed by existing transit infrastructure. This paper proposes a solution that “flips transit on its head.” By rapidly prototyping microtransit services across cities and analyzing supply-demand mismatches, it is possible to launch truly data-driven transit services. To illustrate the framework, a unique dataset generated from a year of Dallas Area Rapid Transit’s GoLink service, one of the largest on-demand microtransit services in North America, is used. Mapping and machine learning are combined to empower planners to “join the dots” when (re)designing fixed-route transit lines. It is shown that microtransit should not simply fill in the gaps left by inefficiently scheduled bus routes: by incorporating it fully into their planning processes, cities and transit agencies could dramatically reverse the fortunes of public transit.
Milk-run routing and scheduling subject to different pick-up/delivery profiles and congestion-avoidance constraints
