Abstract: Lane-changing behavior is one of the basic driving behaviors and a basic algorithm of traffic simulation. It is of great significance to study lane-changing behavior. At present, there is a strong assumption in many lane-changing models established by scholars, that is, all drivers are completely rational, which cant reflect the drivers bounded rationality in evaluating the effectiveness of lanes and makes the model inconsistent with the actual situation. In view of this, the prospect theory was introduced to establish a new lane-changing decision-making model, which could reflect the potential irrational behavior of drivers in the process of lane -changing decision-making. The established model could reflect more realistic driving behavior. According to current research, the two indicators such as front vehicle speed and headway distance were selected to evaluate the drivers satisfaction with the lane. Considering the drivers risk preference and the trade-off of objective probability, the foreground values of different lanes were calculated, and the Logit model was established. Finally, the probability of the driver making a lane-changing decision was obtained. The study found that the proposed model has better accuracy and stronger applicability than the model based on random utility theory does, and the drivers received information and cognition of the behild vehicle on the target lane are relatively limited during the lane-changing decision process.

Key words: traffic and transportation engineering; lane-changing decision; bounded rationality; prospect theory; discrete choice

摘要： 换道行为是基本的驾驶行为之一，是交通仿真的基本算法，研究换道行为有着重要的意义。目前学者们建立的很多换道模型中有一个很强的假设，即所有的驾驶员是完全理性的，无法体现驾驶员在评价车道效用的有限理性，使得模型与实际情况不符。鉴于此，引入前景理论建立新的换道决策模型，可体现驾驶员在换道决策过程中潜在的非理性行为，模型可反映更加真实的驾驶行为。根据现有研究，选用前车速度和车头间距2个指标评价驾驶员对车道的满意度，综合考虑驾驶员的风险偏好以及对客观概率的权衡，在此基础上计算不同车道的前景值并建立Logit模型，最终得到驾驶员做出换道决策的概率。研究发现该模型比基于随机效用理论的模型有更好的准确度，模型适用性更强，而且换道决策过程中驾驶员对于目标车道后车的接收信息和认知比较有限。

关键词: 交通运输工程； 换道决策；有限理性；前景理论；离散选择