Modelling Interurban Multi-modal Trips Using a Fizzy Logic Approach
HOOGENDOORN S, HOOGENDOORN-LANSER S and BOVY P H L, Delft University of Technology, The Netherlands
The future multi-modal transport system will be an integrated mixed, flexible multi- layered network of various types and forms of transportation services, which are linked together in transfer nodes. It includes both individual transport modes (walk, bic
The future multi-modal transport system will be an integrated mixed, flexible multi- layered network of various types and forms of transportation services, which are linked together in transfer nodes. It includes both individual transport modes (walk, bicycle, private car and taxi) and public transport modes (bus, train and subway) offered by private or public suppliers. These public transport modes are organised as a combination of fixed and flexible demand-responsive services (taxi, para-transit). This new system aims to adequately cope with the partly unpredictable time-varying access restrictions, quality of transport supply and economy of fares.
To design a multi-modal transport system and assess the market potential of the transport modes, insight into travellers' appraisal of trip chain attributes is of dominant importance. In principle, behavioural travel choice models can yield such insights. However the use of traditional models is limited, since they rarely consider multi-modal trip chains between origins and destinations.
A multi-modal trip chain consists of legs with different transport modes. Between these legs a transfer movement is necessary. In addition to these between-mode transfers, such as bus-train and train-subway, within-mode transfers, like regional bus- city bus, can be distinguished.
Several mode-specific attributes, such as in-vehicle time and waiting time at transfer points, can be identified. Further, we can distinguish non-mode-specific attributes, Flke the number of transfers. We envisage that the predictive performance of travel choice models can be improved significantly by distinguishing mode-specific trip chain attributes and taking into account typical chain characteristics, such as transfer locations and composition of the chain.
The numerous mode-specific and non-mode-specific trip chain attributes influence travel choice behaviour to some extent during the different stages of the choice process. A hierarchy, closely resembling the actual choice process and accounting for the relevant trip chain attributes at each stage of the choice process, will further improve the predictive ability of travel choice models.
Association for European Transport