A multimodal public transit network design method based on hub-and-spoke infrastructure

The quality of public transport in a modern city depends on the interaction of the different modes of transport. The efficiency of a multimodal public transport network (MPTN) can be improved by adapting the hub-and-spoke model. In this thesis, a two-phase method is presented to generate a hub-and-spoke network for MPTNs. In the first phase, the problem of clustering in MPTNs is addressed. A spatial clustering algorithm is presented that can process real urban MPTNs without reducing their size or dividing them into zones. The algorithm is tested on four large city datasets and compared with the popular spatial clustering algorithms. The largest test network (Sydney, 24063 nodes) was processed in less than a minute, and all clusters generated by the algorithm contain less than 100 nodes, while the clusters generated by the compared algorithms contain significantly more nodes. In the second phase, an algorithm based on multi-criteria decision making is presented, which simultaneously locates hubs and hub lines. The proposed method is tested with the MPTN of Greater London and the resulting hub-and-spoke network is compared with the Journey API provided by Transport for London. The results show that the total travel time is improved by 16.90% with a strict hubbing policy. Key Words: Multimodal Public Transit Network Design, Spatial Clustering, Hub Location Problem, Hub Line Location Problem, Hub-and-spoke network