Regret-based Multi-modal Traffic Assignment Considering Traveler Familiarity

To account for bounded rationality in traveler choice under uncertainty, this study develops a multimodal traffic assignment model grounded in regret theory. The model uniquely incorporates traveler path familiarity, capturing how regret intensity varies across different modes and user experiences. We formulate a random user equilibrium condition that considers heterogeneous familiarity levels from a regret-aversion perspective. An equivalent variational inequality (VI) formulation is proposed, and the existence of its solution is rigorously proven. Numerical results reveal that disparities in route familiarity, inter-modal interactions, and users' psychological aversion to regret significantly impact network equilibrium flow patterns and modal split. Comparative studies on the Sioux Falls network reveal that traditional models may overestimate private car shares by neglecting familiarity-induced regret aversion.The proposed model provides a more behaviorally realistic framework for traffic analysis and offers a robust theoretical basis for designing more targeted multimodal traffic management strategies.