Modelling Peak Spreading in Continuous Time
VAN WREN T, CARMICHAEL S, Hague Consulting Group, POLAK J, Imperial College, HYMAN G, Department of the Environment, Transport and the Regions and CROSS S, Halcrow Fox, UK
In the 1994 SACTRA report changing departure time was recognised as an important response to both increasing congestion and transport policy, second only to routing changes. Unfortunately, techniques available to model departure time choice and peak sprea
In the 1994 SACTRA report changing departure time was recognised as an important response to both increasing congestion and transport policy, second only to routing changes. Unfortunately, techniques available to model departure time choice and peak spreading have suffered from conceptual drawbacks, generally related to the time dimension. For example, many of the existing techniques have simulated choice of departure period (peak/off-peak, or peak hodshoulder), whilst in reality at least a proportion of departure time shifts may be expected to have much smaller orders of magnitude, and to take place in continuous time.
Techniques have been developed in the past to model peak spreading in continuous time for a single bottleneck. Such methods assume personal preferred arrival times (PATS) dependent on, say, trip purpose and personal characteristics, and scheduling trade-offs between travel duration, early and late arrival. This paper presents the implementation of these ideas into a framework of equilibrium scheduling theory (EST), 'allowing network-wide application by generalising individual choices into more aggregate relationships at origin-destination level in continuous time. Software has been developed to interface with standard commercial assignment software (SATURN, TRIPS, CONTRAM), ready for application in practice.
Association for European Transport