A risk-averse user equilibrium model for route choice problem in signal-controlled networks

This paper proposes a new risk-averse user equilibrium (RAUE) model to estimate the distribution of traffic flows over road networks with taking account the effects of accident risks due to the conflicting traffic flows (left- and right-turning and through traffic flows) at signalized intersections. It is assumed in the proposed model that drivers consider simultaneously both the travel time and accident risk in their route choices. The accident risk of a route is measured by the potential accident rate on that route. The RAUE conditions are formulated as an equivalent path-based variational inequality problem which can be solved by a path-based solution algorithm. It is shown that the traditional user equilibrium (UE) model is in fact a special case of the proposed model. A numerical example on a grid network is used to illustrate the application of the proposed model and to compare the results with the conventional UE traffic assignment. Numerical results show that the traditional UE model may underestimate the total system travel time and overestimate the system accident rate. Sensitivity tests are also carried out to assess the effects of drivers' preferences, signal control parameters (i.e., green time proportions), and various network demand levels on the route choice problem.