Rationally Inattentive Route Choice by Link-Based Segments

Abstract

In this paper, we consider a rational inattentive (RI) route choice problem. Since a computational challenge, called curse of dimensionality, always arises in the existing route-based RI choice models. We establish a RI choice model by link-based segments, where the RI choice behaviour is learnt along the links instead of the routes. We introduce the background information into the RI choice model, which ensures its optimal solution locates within the interior of the feasible region. We then analytically characterize the closed-form expression of the optimal solution, which resembles a nested logit model. Finally, a numerical example is illustrated, showing that the solution algorithm can identify the RI route choice behaviour within a reasonable computation timeframe.