Requirements for traffic assignment models for strategic transport planning: A critical assessment

Abstract

Transport planning models are used all over the world to assist in the decision making regarding investments in infrastructure and transport services. Traffic assignment is one of the key components of transport models, which relate travel demand to infrastructure supply, by simulating (future) route choices and network conditions, resulting in traffic flows, congestion, travel times, and emissions. Cost benefit analyses rely on outcomes of such models, and since very large monetary investments are at stake, these outcomes should be as accurate and reliable as possible. However, the vast majority of strategic transport models still use traditional static traffic assignment procedures with travel time functions in which traffic flow can exceed capacity, delays are predicted in the wrong locations, and intersections are not properly handled. On the other hand, microscopic dynamic traffic simulation models can simulate traffic very realistically, but are not able to deal with very large networks and may not have the capability of providing robust results for scenario analysis. In this paper we discuss and identify the important characteristics of traffic assignment models for transport planning. We propose a modelling framework in which the traffic assignment model exhibits a good balance between traffic flow realism, robustness, consistency, accountability, and ease of use. Furthermore, case studies on several large networks of Dutch and Australian cities will be presented.