Bounded-rationality based day-to-day evolution model for travel behavior analysis of urban railway network

Abstract

Existing day-to-day traffic assignment models are all built to capture day-to-day traffic fluctuations, but most of the evolution process itself and the final equilibrium state are based on the assumption of passengers’ rational behavior, that is, to find the path with the minimum travel cost, which ignore the correlation among the days’ evolution and boundedly rational (BR) of travelers in the path choice and thus can give very unreasonable results for ones with this behavior. Such an assumption basically ignores the correlation among day-to-day evolution and bounded rationality (BR) of travelers in the path choice and thus could result in inaccurate results for the travelers with such behavior. This paper proposes a day-to-day dynamic evolution model with the consideration of BR, which can better captures travelers’ characteristics in the path finding within an urban railway network. In order to capture the correlation of path choice over time, we introduce a time series method, detrended fluctuation analysis (DFA), to analyze the complex long-term correlations hiding in the passengers’ evolution over time. The study results clearly show that the proposed model and analytical approach is better for capture the day-to-day dynamics in travel behaviors and can serve as a general framework of modeling passengers’ BR behavior.