Modeling context-dependent decision rules in route choice behavior under uncertain conditions

Motivated by the research on heterogeneity in human choice behavior, we collected data to capture the role of uncertain conditions and trip-related context in route choice behavior of travelers. The model developed considers route choice as the outcome of the simultaneous presence of arrival time variability and multiple decision context effects: activity type, number of people involved and kind of relationship with these people. This study estimates a mixed random utility-random regret structure to reveal potential linkages between characteristics of the trip and respondents' adoption of a particular decision rule using stated choice data. The utility maximization model assumes that choice behavior is based on tastes only. In contrast, the regret minimization model assumes that choices are based on a comparison of choice alternatives. In an effort to better understand and model human behavior, several researchers have explored which of these models better fits a given dataset, but not much research has been conducted to mix these two approaches. Moreover, variability in human behavior may result in an individual exhibiting diverse decision strategies. Individuals may adopt a different decision rule, depending on choice context. To allow for this situation, a latent class model is formulated and estimated. To overcome the limiting assumption of constant membership probability, we allow for behavioral flexibility by introducing activity context as explanatory variables. The findings offer important further insights, not only into the behavioral patterns of respondents, but also into the potential drivers of decision rules.