摘要
为了提升交通分配算法的速度以适应现实业务需求。本文提出了一种改进Frank-Wolfe算法,该算法模型相较于传统模型,优化了道路路阻计算方法,使模型路径规划接近现实情况,并预设出行路径集,避免穷举网络所有路径。此外,本文还解决了路段途径车流OD无法溯源的问题。在测试实验中,本文以深圳市的交通数据为实验数据,用本文的方法进行OD溯源和交通分配,本文中的算法在运行速度上明显优于传统算法,更有利于部署在城市交通管理实际应用上。
In order to improve the speed of the traffic allocation algorithm to adapt to the actual business needs,this paper proposes an improved Frank-Wolfe algorithm.Compared with the traditional model,the algorithm model optimizes the road resistance calculation method,makes the model path planning close to the reality,and presets the travel path set to avoid exhausting all paths in the network.In addition,this paper also solves the problem that the OD of the traffic flow on the road section cannot be traced to the source.In the test experiment,this paper uses the traffic data of Shenzhen as the experimental data,and uses the method of this paper to carry out OD traceability and traffic distribution.The algorithm in this paper is significantly better than traditional algorithms in operating speed,and is more conducive to deployment in urban traffic management application.
作者
郑晏群
张鹍鹏
ZHENG Yan-qun;ZHANG Kun-peng(Shenzhen Transportation Operation Command Center,Shenzhen 518040,China;Shenzhen Transportation Design&Research Institute,Shenzhen 518003,China)
出处
《价值工程》
2021年第14期193-196,共4页
Value Engineering
作者简介
郑晏群(1984-),女,山东临沂人,高级工程师(交通信息工程及控制),硕士研究生,研究方向为智能交通;张鹍鹏(1981-),男,河南鹤壁人,教授级高级工程师(交通信息工程及控制),博士研究生,研究方向为智能交通。
