基于粒子群优化算法的交通数据流聚类分析

针对交通数据流聚类分析过程中生成顺序的不确定性,提出了采用基于网格和密度的D-Stream算法对交通数据流进行聚类分析,并将粒子群优化算法引入聚类过程,从而对数据流聚类分析方法进行了改进,使数据聚类能够根据本身的密度极大值有序生成,增强了用户对聚类过程的控制能力.通过昆明市实测交通数据流进行聚类分析,得到了能够反映交通状况不同特征的聚类结果和动态的控制策略,并对交通数据流的相关研究工作提供决策支持.

Clustering Analysis of Traffic Data Stream Based on Particle Swarm Optimization Algorithm

Aiming at the uncertainty of generating sequence in the traffic stream clustering analyzing process, this paper is employed D-Stream algorithm, which is based on grid and density, to cluster traffic. And hybrid particle swarm optimization algorithm is introduced into the clustering. Then data stream clustering method is improved. It can make clustering generated in order according to the density maximum value itself. It strengthen the users to control the clustering process. At last it obtains some categories that are able to reflect the different traffic status and dynamic control tactics by analyzing the practical traffic data stream of Kunming. Then the result could offer the decision-making support to the related study of traffic data stream.