General two-level framework for demand-responsive transport optimization

This paper proposes a generalized optimization framework for demand-responsive transport (DRT), a modern form of transport organization in which services are planned and provided directly in response to the client demand. The developed framework has a two-level structure. The first level includes basic features essential to every DRT system. The second level takes into account the electric vehicle energy consumption and charging, predefined initial parts of vehicle routes for combining consecutive vehicle planning windows, as well as special passenger requests, such as a dedicated space for disabled people or Wi-Fi network availability. Three variants of models have been prepared, differing in how vehicles move between bus stops: freely point-to-point planning (General), restricted sections of the vehicle path (Sections) and fixed routes of every vehicle (Routes). During experiments, the developed model was thoroughly tested, in particular, its effectiveness in the basic version and with additional extensions was evaluated. The tests were performed on a dataset created based on the real public transport system of Rzeszow, Poland, from which 64 bus stops were mapped. The optimization process involved 200 passengers. The experiments confirmed the usefulness of the proposed solution.