Study On Heterogeneity Of Travelers’ Route Choice Rules Based On Structural Equation Modelling

It was typically assumed in traditional route choice researches that all travellers involved in the researches make route choice according to one same rule. However, the assumption is obviously not in accordance with travellers’real choice behaviors when differences in decision context and travellers’heterogeneity are under consideration. Consequently, heterogeneity of route choice rules is the core concern of this research which is based on the shoulders of the existing researches both from transportation science, decision theory and attribution theory in psychology.Three fundamental assumptions were made in this research. First, travellers has differential preference for route choice rules; second, socio-demographics have effects on the preference for the rules; third, purpose of the trip has some effect on the preference for the rules. To validate these assumptions, four commonly studied rules(i.e., shortest path, acceptable route, familiar route as well as reference dependence, later, at convenience, shortest rule, acceptable rule, familiar rule, reference rule respectively) were included and discussed in detail, and two purposes of trip were sets(i.e., drive-for-work and drive-for-entertainment). In the following part, the concept of route choice was given according to existing researches; factors that may having effects on route choice were clarified into two aspects according to attribution theory in psychology.Then, SEM was introduced in briefly; a initial structural equation model about six socio-demographic variables and four latent variables of preferences for the four route choice rules was developed; two questionnaires were respectively designed for drive-for-work purpose and drive-for-entertainment purpose. After that, the goodness-of-fit of the initial model was given and discussed; the final model and its estimation results were presented.Results from descriptive analysis and the final SEM estimation have demonstrated four interesting findings. First, overwhelming majority of respondents show great preference for more than one rule. Second, the four rules are not mutually independent. Preference rule has significant positive effects on all the other rules. Third, socio-demographic variables indeed have significant effects on the preference for the four route choice rules. Fourth, purpose of trip has effect on the preference for the four route choice rules..To classify drivers into different groups by rule-preference heterogeneity demonstrated in this research shed light on individual-heterogeneity-based multi-class traffic assignment modelling.