C-Weibit Discrete Choice Model: A Path Based Approach

In traffic assignment a key role is played by the path choice that simulate the path chosen by users in relation to the perceived costs. Various models are proposed in literature in order to accomplish this task (e.g., those obtained from generalized extreme value model). One of the main problems encountered is the overlap among paths between the same origin-destination pair. To overcome this issue, the research was directed towards models like C-logit, Path-Size Logit or Probit capable of taking these aspects into account. In this paper the use of a C-Weibit model is proposed, by introducing a commonality factor that modify the utility function when the paths overlap. Among the advantages of this model, it should be noted that the probability can be calculated in closed form. The C-Weibit was used within a stochastic user equilibriumassignment procedure. The results and some comparisons are reported, in these tests C-Weibit had better performances than other models.