This paper reports the outcome of a project for the UK Department for Transport, which forms part of its ongoing research programme on journey time variability (JTV). This research is currently evaluating alternative frameworks for supply models of JTV, including models which consider incident related variability separately from other sources of variability. The paper also reviews alternative methods for incorporating the uncertainty of travel time into utility functions. Consideration is given to the need to reconcile measured and perceived versions of JTV and the practicality of avoiding an incidents/events segmentation, which depends on arbitrary definitions. Existing models including the INCIBEN/INCA models developed earlier in the ITEA research programme, and an incident variability model are evaluated. Finally, the paper develops generic models of JTV and considers the detailed problems faced in using automatic number plate recognition (ANPR) data collected on UK motorways to calibrate such models. For the covering abstract see ITRD E124693.
Abstract