Research On Repeated Route Choice Behaviour Based On Information Reliability

Traffic information services can uniformly distribute traffic flow over the traffic network to alleviate traffic congestion by changing travelers' behavior.Whether its objective can be achieved or not depends on travelers' response.In terms of information content,traffic information that affects route choice can be categorized as descriptive information and prescriptive information.Due to the uncertainty of traffic demand,traffic incidents and severe weather,the road status information often is different from the real status and unreliable.Therefore,considering travel experience,this dissertation conducts a research about repeated route choice behavior based on information reliability to explore the effect of two types of unreliable information on repeated route choice behavior and on traffic flow,which provides support for the management measure design and evaluation of traffic systems with traffic information.The main content of this dissertation can be divided into four parts including investigation about the effect of information reliability on repeated route choice;exploring travelers' information reliability perception mechanism;the effect of information reliability on repeated route choice and the effect of unreliable information on traffic flow.(1)Investigation about the effect of information reliability on repeated route choice.The laboratory experiment with low cost and outcome feedback is chosen after comparing three existing data acquisition methods.Based on design principles of the laboratory experiment and the purpose of the investigation,we design an experiment and establish an experiment system by the Java Script language under the framework of vue.The system helps us collect data about the perceived information reliability and repeated route choice.(2)Exploring travelers' information reliability perception mechanism.Considering information use experience,we establish several perceived information reliability models based on cognitive psychology theory and ordinal regression model.The nonparametric test is then adopted to find the factors that affect the evaluation of information reliability.These factors are further incorporated into the information reliability perception models.The fitting results reveal travelers perceive descriptive information reliability by learning descriptive information reliability experience and prescriptive information reliability based on current experience.(3)The effect of information reliability on repeated route choice.The nonparametric test is used to verify that there is an effect of information reliability on information use.Based on the discrete choice model and travel time learning model,a logit route choice model and a mixed logit route choice model are established to qualify the effect of information reliability on information use and to disclose travelers' repeated route choice mechanism,which can provide insights for travel information services.The results show that the difference between travelers' repeated route choice behaviour is insignificant in unreliable information scenario;the reliability of the information affects travelers' route choice by affecting the information use;the reliability of the information has a positive relationship with the information use.(4)The effect of unreliable information on traffic flow.Considering the perception and effect of information reliability,we,respectively,construct day-to-day traffic assignment models with descriptive information,prescriptive information and two types of information to provide methods for analyzing the effect of unreliable information on traffic flow.We then explore the natures of these models through non-linear dynamics theory and numerical experiments to demonstrate traffic flow evolutionary mechanism,which provides support for the management measure design of traffic systems with traffic information.We also propose some ways to stabilize traffic flow.Finally,we analyze the effect of two types of unreliable information on traffic flow of a typical region to show the application of the model.