Hybrid rational route choice approaches: Using concepts from fuzzy logic and the analytic hierarchy process

This research work represents two route choice models based on fuzzy logic and approximate reasoning for explaining route choice behavior in transportation planning. The first model is based on Weber's psycho-physical law of 1834. In this model, a set of fuzzy ‘if-then’ rules is developed to represent a typical driver's psychology for capturing preferences, pairwise, among alternatives that a person may consider. The second model is based on Teodorovic and Kikuchi's (1990) fuzzy ‘if-then’ rules. The Analytical Hierarchy Process (AHP) is incorporated in both models, as part of the second stage of the models to represent the underlying decision-making mechanism. The advantage of the AHP is that it offers the flexibility for developing fuzzy ‘if-then’ rules and its theoretically justified scale for purposes of comparison.; First, route choice behavior is explained hierarchically for these models. Then, factors such as travel time, safety and congestion in route choice decision-making are modeled as fuzzy numbers. Based on the methodological features of these models, the input domains are categorized and linguistically labeled. Using fuzzy logic and approximate reasoning, drivers' preference allotment among the alternatives are assessed, pairwise. The results obtained from this first stage are then used as inputs for the AHP's pairwise matrices. Drivers' preferences on each alternative are estimated considering only one factor at a time. The final preference allocation among the alternatives is derived using the least squared error optimization technique.; Finally, these models are applied to a specific network to evaluate their effectiveness. These models are able to explain drivers' route choice behavior at both disaggregate and aggregate level with statistical significance. At the aggregate level, we also tested the predictive ability of these models against the traditional logit model. The results from these models provide better fit than the results obtained from the traditional approach.; These models provided that drivers' decision making process can be explicitly explained by a small set of intuitive and reasonable fuzzy ‘if-then’ rules and by the AHP. The results obtained show that these models present promising mathematical approaches with the ideas from psychology to model efficiently complex drivers' route choice decision making process in transportation planning.