Travel decision making under uncertainty and road traffic behavior: The multifold role of ambiguity attitude

To connect commuters’ uncertain mode choices with traffic behavior in the presence of travel time variability, we develop a dynamic traffic simulation in terms of an agent-based model, which consists of two sub-models, the mode choice model and the traffic flow simulation model. The modeling framework accommodates the interplay between mode choice and traffic behaviors and their co-evolution over time. We embed an extended list of empirical parameters including ambiguity/risk attitudes and time-money trade-offs within a rank-dependent and mode-dependent utility framework to imitate commuters’ daily mode choice behaviors. The improved behavioral realism at the micro-level results in more realistic outputs such as modal split and average speed at equilibrium. The evidence demonstrates that a richer representation of mode choice behavior at the individual level is associated with a gain in aggregating them to the level of system behavior, in which ambiguity seeking, a typical behavior in the loss domain but largely ignored in the transport literature, plays an important role.