Modeling Departure Time Choice For Shopping Trips Based On Cumulative Prospect Theory

The shopping trip is the second important travel activities only to the commute trip. With the economic and social development, consumer demand, followed by shopping trip’s traffic volume increases. The congestion appears on weekends and the traffic flow is different from commuter days. Therefore it’s necessary to study the shopping trip, for providing the basis for the formulation of traffic control measures.More and more traffic information will be studied by travelers to abstract more effective information for decision-making, as well as Advanced Traveler Information System is building. Accordingly, the estimation of travel time approach reality gradually that means precision of time estimation is augmenting. Though so much information travelers could acquire, they still confront of uncertainty and risk for occurrence will happen is unknown.Expected utility theory (EUT) is used to analyze decision under risk. But empirical research has shown that it fails to provide a good description of individual behavior in situations of uncertainty. Building upon its predecessor prospect theory, cumulative prospect theory (CPT) which introduces empirical research, has become one of the most prominent of decision-making of uncertainty nowadays. This paper compares EUT and Random utility theory (RUT) with CPT by qualification, postulates, conditions, methods, decision and process.In this paper, the characteristics of the shopping trip are analyzed and departure time choice models have been built. In editing phase, reference points are assumed firstly. Secondly, value functions with arrive time value and trip time value are built. Thirdly, individual subjective probability is deducted for calculating weights. In evaluation phase, commuters and shoppers choose the option with the largest expectation. Thereby a parameter is used to estimate individual subjective congest degree, which shows individual ability to combine experience with new traffic information. Besides, Bayesian theorem is consult to build dynamic model for adjust departure time.Finally, a small-scale travel behavior survey is taken to analyze trip activities and demarcate parameters for forecasting departure time is calibrated. After calculation model, a shopping trip departure time distribution diagram is drawn to predict the shopping peak hours and analysis congestion. IF path selection is considerate, times-network traffic flow diagram which provides additional reference information will be shown for traffic managers to develop measures.