Abstract
Most route choice models describe route choice behavior in habitual networks, where drivers have spatial knowledge of networks based on environmental experiences. However in disaster networks, drivers do not have any experiences or appropriate information. As a result, drivers choose routes based on dynamic decision-making, which is myopic and somewhat forwardlooking. In order to describe this decision-making dynamics, we proposed a sequential route choice model with a sequential discount rate, which is a discount factor of expected future utility. Through illustrative examples, we showed that the sequential discount rate re?ected drivers decision-making dynamics and had a large effect on evaluation of route choice behavior. We also estimated the parameter of the proposed model using the data collected in the Tokyo Metropolitan area on March 4 and 11, when the Great East Japan Earthquake occurred. The results showed that the sequential discount rate had a lower value in gridlock networks than ordinary networks and route choice mechanisms could change dynamically.
