Transit Voronoi diagrams in multi-mode public transport networks

Abstract

Network Voronoi diagrams (N-VDs) are effective geometric constructions for partitioning geographical space constrained by road networks. However, they are not applicable for urban areas with well-developed public transport systems. This study proposes new transit Voronoi diagram (T-VD) models for partitioning geographical space constrained by public transport networks. The proposed T-VD models explicitly consider the complexities of a public transport network, including transfers between different transport modes, dynamic transit schedules, changing network topologies, etc. To efficiently construct T-VDs, geo-computational algorithms are developed by modifying and integrating classical shortest-path algorithms in road networks and transit shortest-path algorithms in public transport networks. Case study results show significant differences between T-VDs and N-VDs, highlighting the need for using T-VDs in urban areas with well-developed public transport systems. The developed geo-computational algorithms efficiently constructed T-VDs in large-scale public transport networks within satisfactory computational times.