Bayesian Networks for travel demand generation: An application to Switzerland

Abstract

Bayesian Networks (BNs) are probabilistic graphical models representing conditional dependencies existing between variables of interest. Recent studies have employed BNs for population synthesis and daily activity plan generation. Those studies highlight the ability of BNs to efficiently detect the causality links between variables in an easily interpretable way. This short paper aims to propose a further application of BNs for both population and daily activity plan synthesis in Switzerland. We show that understanding the dependency structure linking the population characteristics and its mobility behaviour is key to generating representative synthetic activity patterns. Furthermore, we lay the foundations for the development of temporally transferable travel demand models.