购物出行与交通方式组合出行链仿真分析

发布时间：2019-04-12 14:57

【摘要】：居民交通出行活动研究是制定城市交通政策和规划的重要基础。考虑了出行活动一系列的选择过程,将出行活动和交通方式按照发生序列创建组合链,组合链包含了出行目的、交通方式、活动序列等信息。通过组合链可以得到购物出行频率和交通方式指标。以交通出行需求条件弹性较大的购物出行为基础,利用居民一日出行活动日记数据,提取购物出行组合链。利用神经网络并行处理及非线性拟合的优势构建组合出行链预测模型,分别建立了基于传统神经网络、LM神经网络和切比雪夫神经网络的仿真模型,输入出行者社会经济属性、家庭特征等17个影响因素对组合链模型进行训练,经验证切比雪夫模型的预测效果最优。标定后的模型可以较好的预测购物出行频率和交通方式选择,可为交通需求管理政策研究、效果预判等提供量化分析工具。
[Abstract]:The study of residents' traffic travel activities is an important basis for the formulation of urban traffic policies and planning. Considering a series of selection processes of travel activities, a combination chain is created according to the occurrence sequence of travel activities and traffic modes. The combined chains contain information such as travel purpose, traffic mode, activity sequence and so on. Through the combination chain, we can get the shopping travel frequency and the traffic mode index. Based on the shopping trip, which has great elasticity in the demand condition of traffic travel, the shopping trip combination chain is extracted by using the daily travel activity diary data of residents. Making use of the advantages of parallel processing and nonlinear fitting of neural network, the prediction model of combined departure chain is constructed. The simulation models based on traditional neural network, LM neural network and Chebyshev neural network are established respectively, and the socio-economic attributes of travelers are inputted. The combination chain model was trained by 17 factors, such as family characteristics, and the prediction effect of Chebyshev model was proved to be the best. The calibrated model can predict the frequency of shopping trip and the choice of traffic mode. It can provide a quantitative analysis tool for the study of traffic demand management policy and the prediction of effect.
【作者单位】：北京交通大学城市交通复杂系统理论与技术教育部重点实验室;北京交通发展研究中心;北京交通大学计算机与信息技术学院;
【基金】：973国家重点基础科学研究项目(2012CB725403) 国家自然科学基金项目(51178032)
【分类号】：U12

【参考文献】