Link-based stochastic loading methods for weibit route choice model

Abstract

The multinomial logit route choice model is known to have some drawbacks in traffic loading because of the independently and identically distributed assumption with the Gumbel distributed random error term. This assumption results in the model's inability to handle the route overlapping and heterogeneous perception variance problems. This study proposed a new traffic loading technique for handling both route overlapping and heterogeneous perception variance problems. Specifically, a modified Dial's STOCH algorithm and a link-based commonality factor were combined with the weibit route choice model to create the link-based C-weibit stochastic loading method. Numerical examples show the features of this stochastic loading method. The Chicago, Illinois, sketch network was used to demonstrate the applicability of the C-weibit stochastic loading method in a real transportation network.