Proactive route choice with real-time information: Learning and effects of network complexity and cognitive load

摘要

Proactive route choice refers to a driver's taking into account future diversion possibilities enabled by real-time information in a network with random travel times. Route choice experiments were conducted in three types of networks with increasing complexity, where the simplest one has no diversion possibilities and is used to gauge risk attitude. Two apparatuses with different cognitive load, a driving simulator and PC were used. Aggregate analyses show that network complexity negatively affects subjects' ratio of choosing the risky option on a driving simulator; cognitive load negatively affects subject's ratio of choosing the risky option in the simplest network, but not in more complex networks. A mixed Logit model with two latent classes, proactive and myopic, is specified and estimated. Results show that subjects learn to be more proactive over time. The impact of network complexity or cognitive load on being proactive is not statistically significant. Cognitive load however increases myopic subjects' risk aversion.