引入土地结构熵参数的出行分布预测广义重力模型研究

Generalized Gravity Model of Land Structure Entropy Parameters for Trip Distribution Prediction

在交通出行分布预测中,一般采用非重力模型的方法单独估计区内出行量,而这会导致与出行分布理论不一致性,以及复杂的统计分析。文中借助信息熵理论,建立了土地结构熵模型和引入熵参数的广义重力模型,以获得区内出行和区间出行一致的理论算法。将城市用地规划的构成要素直接引入交通规划,实现了土地利用与城市交通在规划理论和技术上的衔接,提高交通规划的科学性和合理性。以2008年西安市大规模居民出行调查实际数据为基础,进行了区内出行和整体出行预测的误差分析。验证结果表明:构建的广义重力模型有效提高了区内出行估计精度,预测精度提高1倍,误差降低5个百分点,最大误差降低达到15个百分点。

In trip distribution prediction,a non-gravity method is usually adopted to estimate trip distribution at a traffic zone,which leads to the inconsistency of trip distribution theory and the complicated statistical analysis.Based on the entropy function,this paper develops a land structure entropy model and a generalized gravity model of the entropy parameter so as to get a same theoretical expression of inside-zone trip distribution and inter-zone one.Based on the field survey data in Xi＇an,the error analysis is conducted.The results show that the prediction accuracy is increased by one time,while the error is reduced by 5percent,and the maximum error is reduced to 15 percent.