Integrating Travel Demand and Network Modelling: a Myth or Future of Transport Modelling

Abstract: In this paper, a novel transport planning model system (TPMS) is formulated which is built on the concepts of supernetworks, multi-modality, integrity and calibration. In the proposed formulation, activity travel pattern (ATP) choice facets including the choices of activity, activity sequence, mode, departure time, and parking location, are all unified into a time-dependent supernetwork. The proposed model accounts for the dynamicity of the network, including time-of-day and congestion effects. These help capturing the interdependencies among all different attributes of a full transport planning system. Moreover, the proposed TPMS explicitly formulates an operating capacitated public transport system. To allow visiting locations multiple times and to alleviate the complexity of the proposed supernetwork, a novel multi-visit vehicle routing problem is proposed which does not enumerate the node and link visits. In order to calibrate the model based on the major travel attributes of the travel survey data, a set of splitting ratios are introduced to distribute trips on the supernetwork. The model uses the splitting ratios to integrate the supernetwork and the traffic assignment model in a unified TPMS structure. At last, numerical examples are provided to demonstrate the advantages of the proposed approach.