The static on‐demand bus routing problem: large neighborhood search for a dial‐a‐ride problem with bus station assignment

Abstract

AbstractWe introduce a novel optimization problem to support the planning and routing of on‐demand buses in an urban context. We call this problem the on‐demand bus routing problem (ODBRP). Given are a fleet of buses with fixed capacity, a set of bus stations and travel times between them, and a set of transportation requests. Each transportation request consists of a set of potential departure and a set of potential arrival bus stations, as well as a time window, that is, an earliest departure time and a latest arrival time. The aim of the ODBRP is to (1) assign each passenger to a departure and arrival bus station and (2) develop a set of bus routes to fulfill each request in time while minimizing the total travel time of all users. We present the static version of the ODBRP, as well as a straightforward large neighborhood search heuristic to solve it. The performance of the heuristic is established by comparing it to an off‐the‐shelve heuristic solver (LocalSolver). We also use our heuristic to solve (slightly modified instances of) the well‐known dial‐a‐ride problem. The results found by the heuristic for the on‐demand bus system are compared to those of a simulated traditional public bus system with fixed lines and timetables. A thorough analysis of the comparison demonstrates that total user ride times can be significantly lower in an on‐demand public bus system and shows that an on‐demand bus system works best with a large number of small buses.