Disaggregate route choice models based on driver learning patterns and network experience

Since their emergence, route choice models have been continuously evolving; particularly because of their wide application and consequent influence in the transportation engineering arena. Although early versions of route choice models were based on theories of rational behavior and neglected limitations of human cognition, later closer observance of human behavior resulted in better modeling frameworks such as Bounded Rationality and Prospect Theory. Nonetheless, recent developments in Intelligent Transportation Systems have increased the demand for more exploration, modeling and validation of behavioral route choice models. This work presents statistical models of route switching based on a real-time driving simulator study of 50 drivers. The research presented in this paper demonstrates that (a) different driver learning patterns have significant route choice effects, (b) driver route choice behavior significantly changes with driver network experience, and (c) disaggregate route choice models based on either driver learning patterns or network experience outperform aggregate route choice models.