Research On Travel Route Choice Behavior Considering Driver Heterogeneity

The travel route choice is a typical problem of travel behavior decision-making.It affects the economics of individual travel decisions by drivers at the micro-level,the rationality of traffic flow distribution on the road network at the meso-level,and the scientificalness of traffic management policy formulation at the macro-level.As the main body of decision-making for travel routes,drivers have significant heterogeneity and bounded rationality due to their individual attributes,psychological characteristics,driving experience,information processing capabilities,etc.,and they cannot simply be regarded as the homogeneous group and be constructed the same criteria of route decision-making when studying their travel route choice behavior.This article aims to organically combine the driver heterogeneity with bounded rationality,explore a new perspective of travel route choice behavior modeling and analysis,and reveal the microscopic mechanism of the driver’s travel route choice behavior in an uncertain traffic travel environment based on the latent class model and cumulative prospect theory.This article mainly includes the following research contents:1.Based on the analysis of the applicability of the latent class model to the driver heterogeneity research,the risk attitude,road anger,and driving habit are considered to set the constructive dimension of driver heterogeneity from the perspectives of the driver’s psychology,emotion,and habits.The existing scales of driving behavior are combined to set up observation indicators to measure the heterogeneity,a heterogeneous group segmentation model of drivers based on the latent class model is constructed,and the influencing factors of latent driver’s class are analyzed through multiple logistic regression to identify the heterogeneity of the behavior for different types of drivers.2.Under the framework of bounded rationality,the psychological heterogeneity reference point of drivers for the expected travel time is set considering the individual heterogeneity.The results of driver heterogeneous group classification are integrated into the value function and probability weight function in the cumulative prospect theory,and the relationship between heterogeneity and bounded rationality is established in the form of specific heterogeneous parameters.The optimal cumulative prospect value of each alternative route is used as the criterion of the travel route decision-making,a travel route decision-making model considering the heterogeneity of drivers is constructed to describe the mechanism of the travel route choice behavior of different types of drivers.3.The parameters of the driver heterogeneous group segmentation model are estimated using the questionnaire survey data,to analyze the behavior heterogeneity of different types of drivers.The data analysis results of the questionnaire survey show that the drivers interviewed can be divided into three types: Class 1(risk-appetite and willful group),Class 2(risk-neutral and stable group),and Class 3(risk-averse and cautious group),and the proportions of these three groups in the total samples are 18.15%,43.08%,and 38.77%.Drivers interviewed in Class 1 have a clear appetite for risk,and they are irritable and easily violated traffic rules during driving;drivers interviewed in Class 2showed a risk-neutral attitude towards risk,and their emotions and behaviors during driving were relatively stable;drivers interviewed in Class 3 showed strict risk aversion,a strong ability to control negative emotions,and good driving habits during driving.In addition,the driver’s gender,age,driving experience,travel purpose,and travel time have a significant impact on the latent class to which they belong.4.The test driver’s response to a specific probability travel situation are captured through the design of heterogeneous preference behavior experiments,and the heterogeneous parameter values corresponding to different types of drivers in the route choice model are estimated.The parameter estimation results show that: The test driver’s risk attitude towards travel time gains or losses and the degree of loss avoidance are not as sensitive as economic decision makers’ gains and losses of wealth;The parameter estimates of different types of drivers are significantly different,From Class 1 to Class 3,the value function parameters α and β of the test drivers gradually decrease,and the λvalue gradually increases,while the probability weight function parameters δ and γgradually decrease,indicating that the test driver’s risk attitude toward travel time,the degree of loss aversion,and the sensitivity to objective probability gradually increase.5.The route choice behavior of test drivers in two specific probabilistic travel scenarios are tested by the design of two rounds of travel route choice behavior experiments,and the experimental results are compared with the prediction results of the route choice model.The results show that: Class 1 test drivers are more inclined to choose high-risk(low probability)and high-yield(large travel time reduction)travel routes,Class2 test drivers are more inclined to choose a travel route with a moderate degree of risk and benefit,and Class 3 test drivers are more inclined to choose a travel route with low risk(high probability)and low profit(smaller travel time reduction).The heterogeneous parameter route choice model that takes the heterogeneity of drivers into account has a prediction accuracy of 88.30% for the test driver’s route choice behavior,which is 16.60%higher than the prediction accuracy of the original parameter model(Tversky and Kahneman).The former has stronger applicability for predicting the route choice behaviors of the test drivers.