路况信息发布时出行者有限理性路径选择行为及建模研究

Study on Bounded Rational Route Choice Behavior and Modelling when Real-time Traffic Information Provided

实时路况信息的提供,有助于出行者对不同路径效用的判断,但由于信息的模糊性以及出行者有限的计算能力,出行者的决策行为在某些情况下依然是有限理性的,即不能选择效用最大路径。因此,本研究基于西蒙的有限理性理论,假设出行者不能完全预测全部备选方案的后果,即实现效用最大化,提出了有限理性模型,并设计试验,验证交通领域路径选择中有限理性的决策行为。试验中,每次出行前向被试者发布实时路况信息,记录被试者的决策结果、决策时间等数据,分析其有限理性路径选择行为,并建立有限理性路径选择模型描述其有限理性特性。研究发现出行者在两条路径效用差小于一定阈值时无法实现理性判断,实现效用最大化,只能随机选择;在效用差大于该阈值时才可进行理性判断,实现效用最大化。根据这一现象,将原有Logit模型效用最大化假设松弛为有限的效用最大化假设,并建立有限理性下的Logit模型(BRBL模型)。通过对模型研究分析,发现模型在0点附近的敏感度要小于传统Logit模型。通过分别对BRBL模型和BL模型进行标定,发现改进后的模型可以解释出行者路径选择决策中的一些有限理性行为,而且标定结果显示模型精度优于传统的BL模型,在路况信息提供下的路径选择问题,新的模型更为适用。

Providing real-time traffic information helps travelers to judge the utility of different routes. However, the decision behavior of traveler is still bounded rationality, sometimes called non optimal choice, in some cases due to the vagueness of information or the limited computing power. Therefore, it is assumed that the traveler cannot fully predict the consequences of all alternatives which means maximizing utility based on Simon＇s theory of bounded rationality, and put forward a bounded rationality model. An experiment to verify the bounded rational decision behavior in traffic route choice is also designed. In the experiment, real - time traffic information is released to the subjects before each trip, the data such as the decision results and decision time of the subjects are recorded after each route choice. The bounded rationality route choice behavior is analyzed, and a bounded rationality route choice model is established to describe the bounded rationality of travelers. It is found that （ 1 ） when the utility difference of the two paths is less than a certain threshold, travelers cannot realize the rational judgment and utility maximization, and only the random decision can be made; （2） when the utility difference is larger than the threshold, the rational judgment can be made and the utility is maximized. According to this phenomenon, the maximization hypothesis of traditional Logit model is relaxed to the limited utility maximization hypothesis, and the Logit model under bounded rationality （bounded rational binary Logit model, BRBL） is established. Through the analysis of the model, it is found that the sensitivity of the model in the vicinity of 0 point is less than that of traditional Logit model. After calibrating BRBL model and BL （binary Logit） model, it is found that the improved model can explain some bounded rational behavior in the decision making of route choice, and the calibration result shows that the accuracy of BRBL model is better than that of traditional BL model, the new model is more suitable for the route choice problem when the road information provided.