基于离散多项logit模型的慢行交通分布预测

Slow Traffic Travel Distribution Forcasting Based on Multinomial Logit Model

慢行交通需求预测是慢行交通规划的基础。通过分析慢行出行影响因素,筛选出行距离、就业吸引、慢行环境以及出行者特征作为主要特征变量。建立多项logit模型,基于随机效用最大化理论,计算慢行小区出行选择概率。输入2016年成都市居民出行调查数据,对特征变量系数进行标定,建立六类不同出行目的下的效用函数。然后以成都市中心城区某片区为例,将该片区划分为13个慢行小区,计算某慢行小区至其它小区的出行概率,并对片区慢行出行生成总量进行分布预测,得到各个慢行小区的出行分布量。最后,结合片区土地利用性质以及实际调查,验证所提出的预测方法的实践价值以及可行性。

Slow traffic travel demand forecasting is the basis of slow traffic planning. By analyzing the influencing factors of slow travel, the travel distance, employment attraction, slow travel environment and traveler characteristics are selected as the main characteristic variables. Multinomial logit model are established to calculate the travel selection probability of the slow travel zone based on the random utility maximization theory. Taking the 2016 Chengdu resident travel survey data as input, the model parameters are estimated, and six travel utility functions under different travel destinations are established. Then take a certain area in the downtown area of Chengdu as an example, divide the area into 13 slow travel zones, then calculate the travel probabilities of a slow travel zone to other zones, and predict the slow traffic travel distribution of the area. The travel distribution amount of each slow travel zone is obtained. Finally, through the land use and field survey analysis of the area, the practical value and feasibility of the prediction method proposed in this paper are verified.