摘要
针对传统方法存在的缺点,提出一种基于粒子群K均值聚类算法的电梯交通模式识别方法.该方法通过对此前一周的原始客流数据进行聚类分析,得到相应交通模式的聚类中心坐标.针对实时变化的交通流数据,采集5min时段客流数据,根据最近邻原则划分其归属的聚类中心,从而识别出当前的交通模式.仿真实验表明,该方法能对电梯交通模式进行有效识别,实时性较好.
To overcome the drawbacks of traditional methods, a method of elevator traffic pattern recognition based on particle swarm optimization K-means clustering algorithm is proposed.The traffic flow data during latest 7 days as a sample is applied to clustering analysis, and the clustering centers of the corresponding traffic patterns are obtained by using this method.Five minutes traffic flow data are real-time collected and its corresponding clustering center is partitioned according to the closest neighbor principle, and then the current traffic pattern is recognized.Simulation shows that the proposed method can identify elevator traffic patterns effectively with good real-time performance.
出处
《控制与决策》
EI
CSCD
北大核心
2007年第10期1139-1142,共4页
Control and Decision
基金
国家自然科学基金项目(69975013)
关键词
电梯交通模式
粒子群K均值聚类算法
电梯群控系统
Elevator traffic pattern
Particle swarm optimization K-means clustering algorithm
Elevator group control system
