基于双层规划的老城区自行车专用道网络设计  

Design of bicycle lane network in China’s old city based on bi-level programming

作　　者：王丰 刘昊翔 WANG Feng;LIU Hao-xiang(School of Automotive and Transportation Engineering,Hefei University of Technology,Hefei 230009,China;Engineering Research Center for Intelligent Transportation and Cooperative Vehicle-Infrastructure of Anhui Province,Hefei University of Technology,Hefei 230009,China)

机构地区：[1]合肥工业大学,汽车与交通工程学院,合肥230009 [2]合肥工业大学,安徽省智慧交通车路协同工程研究中心,合肥230009

出　　处：《交通运输工程与信息学报》2022年第4期72-87,共16页Journal of Transportation Engineering and Information

基　　金：国家自然科学基金项目(71801067,72188101);中国博士后基金特别资助项目(2021T140173)。

摘　　要：自行车作为一种慢行交通方式,其基础配套设施的规划是城市交通规划领域的重要组成部分。本文在当前的自行车网络设计研究基础上,针对国内城市老城区需求多、道路密、机非混合的特点,提出了一个新的自行车网络设计问题,即采用新建自行车专用道和限行机动车的方式构建连通的自行车交通网络,旨在提供更为便利的自行车基础设施,以达到促进自行车出行和缓解交通拥堵的目的。本文构建了一个双层规划模型,上层为政府新建自行车道或禁行机动车的决策模型,包括自行车道网络的连通性约束;下层为考虑用户出行需求弹性的Logit模型和路径选择行为的用户均衡(UE)交通分配模型。提出了改进的遗传算法对构建的双层规划模型进行求解,并嵌入Frank-Wolfe算法和自适应平均算法分别求解UE交通分配模型和需求划分的Logit模型。最后,采用Sioux Falls网络进行数值算例分析,结果显示本文的模型和算法能有效地求得连通的自行车专用道网络设计方案,并降低路网的出行总阻抗。As slow transportation,bicycle infrastructure planning is integralto urban transportation planning.Based onprevious studies on bicycle network design and given the significant demand,dense roads,and mixed traffic flow characteristics in old cities of China,we propose a new bicycle network design problemofconstructing a connected bicycle infrastructure network by building new bicycle lanes and restricting motor vehicles.The study aims to provide a more convenient cycling infrastructure to promote cycling and reduce traffic congestionbydeveloping a bi-level programming model.The upper lever is a decision model for the government to build new bicycle lanes or restrict motor vehiclesand includes the connectivity constraints of the bicycle network.The lower level is the combined Logit model-based mode split and user equilibrium(UE)traffic assignment model,considering the user’s travel demand elasticity and route choice behavior.We further propose an improved genetic algorithm to solve the bi-level programming model was.The Frank-Wolfe and adaptive average algorithms are embedded to solve the UE traffic assignment model and Logit model for mode split,respectively.Finally,the SiouxFalls network is used for numerical analysis.The results show that the proposed model and algorithm can effectively solve the near-optimal connected bike network design plan and reduce the total travel impedance of the road network.

关 键 词：城市交通 自行车专用道网络 双层规划 出行方式划分 遗传算法 

分 类 号：U491[交通运输工程—交通运输规划与管理] U121[交通运输工程—道路与铁道工程]