基于随机系数Logit模型的市内出行方式选择行为研究

出行方式选择行为研究对出行行为分析和预测具有重要意义. 以往研究假定某一群体内的所有出行者的偏好都一样，这与实际不相符. 随机系数Logit 模型假定出行者的偏好不一致，并能分析出行者个体特征对于出行者偏好的影响. 本文利用随机系数Logit 模型对市内机动化出行行为研究. 考虑到心理因素对出行者出行选择行为具有影响，本文在研究中加入了出行可靠性、出行舒适性、出行灵活性3 个潜在心理因素. 通过研究表明，随机系数Logit 模型比传统的离散选择模型具有更高的拟合度. 随机系数Logit 模型的效用函数中，步行时间的系数为非随机变量；但车内时间的系数为随机变量，且与出行者结婚状况，是否有车，是否开车上班，月收入是否超过10 000元，以及对交通灵活性的主观需求相关.

Travel Mode Choice in City Based on Random Parameters Logit Model

There is of great importance of travel mode choice analysis on the analysis and prediction of travelers' travel behavior. However, most of papers related to value of time assumed that travelers' preferences were consistent within a group, which is inconsistent with the fact. The random parameters Logit model allows parameter values to vary across the population, and can analyze the impact of travelers' demographic characteristics on parameters. This paper analyzes the motorized travel in the city with the random parameter Logit model. In consideration of impact of the latent psychological factor on travelers' mode choice, this paper also takes into account of travelers' demand of comfort, reliability and flexibility when traveling. The result shows that the random parameters Logit model fitting the data much better than the traditional discrete choice model. From the result, it can be concluded that, the parameter of walk time is non-random. However, the parameter of in-vehicle time is random, and is affected by travelers' marriage, car-ownership, commute-by-car-or-not, whether the monthly income is more than 10 000 Yuan or not, and the requirement of flexibility when traveling.