Traffic assignment based on individual risk-attitude

Abstract

Incidents whose frequencies are unknown cause a large part of the unreliability of road networks. Especially risk-averse road users avoid the possibility of large travel times under a disruption. This paper proposes a traffic assignment which accounts for risk-averseness. Scenarios for all possible disruptions are simulated. The risk-averseness of a user is generalized in a two-parameter route choice formulation. Depending on the parameters, users assign a weight on the travel times for each scenario. This is the basis for their route choice. The simulator can handle a traffic flow composed from different user classes simultaneously in the same network. In the presented case study, the link loads depend on the parameter settings of the users. The route choice for a user class also depends on the risk-averseness of the other users in the network.