A Bi-Level Optimization Model for Network Design of Demand-Responsive Transit

Demand responsive transit (DRT) has gained increasing attentions over these years. However, DRT routes have not been reasonably designed since the previous DRT route planning mainly depends on the experience of operators, leading to unnecessary detours and low attendance. This study proposed an optimization methodology that would help organizers to better select DRT stations and design the network of DRTs within multiple origins and single destination (“many to single”). We formulate a bi-level model considering the benefit and relationship of transit companies and passengers, which provides an avenue for understanding the trade-offs among conflicting groups. Genetic algorithm is developed to solve the issue for large scale network. A case study is applied to demonstrate the proposed model. The results extend traditional DRT route design within “single to single” to DRT network design within “many to single.”