【摘要】 为缓解城市拥堵情况，提出一种基于改进变色龙(Chameleon)算法的交通控制子区划分方法。综合考虑交叉口间距、交通车流量等因素影响，计算各相邻交叉口的流量关联度，构建相似性矩阵；引入密度峰值聚类算法改进变色龙算法，通过度量公式得到子区划分结果。选取义乌市某区域路网进行模型验证分析，结果表明该方法与常用的固定配时法及谱聚类法相比在平均排队长度上降低7.9%和6.2%,停车次数降低32.6%和16.5%,平均延误时间降低17.8%和11.9%,该划分方法能使城市路网子区划分合理，控制效果显著。更多还原

【Abstract】 To alleviate urban congestion, a traffic control sub-area division method based on the improved Chameleon algorithm was proposed. Taking into account factors such as intersection spacing and traffic flow, the flow correlation between adjacent intersections was calculated and a similarity matrix was constructed. The density peak clustering algorithm was introduced to improve the chameleon algorithm, and the sub-region division results were obtained through a measurement formula. A certain regional road network in Yiwu City was selected for model validation analysis. Results show that this method reduces the average queue length by 7.9% and 6.2%, the number of stops by 32.6% and 16.5%, and the average delay time by 17.8% and 11.9% compared to commonly used fixed timing methods and spectral clustering methods. This partitioning method can make the sub-area division of the urban road network reasonable, and the control effect is significant.更多还原