平面移动式立体车库客流状态聚类研究

兰州交通大学自动化与电气工程学院,甘肃兰州 730070

交通运输工程; 立体车库; 客流状态; 模糊聚类算法; 顾客到达; 服务时间

Passenger flow state clustering in flat mobile automated garage
HE Yunpeng and LI Jianguo

School of Automation and Electrical Engineering, Lanzhou Jiaotong University, Lanzhou 730070, Gansu Province, P.R.China

transport engineering; automated garage; passenger flow state; fuzzy c-means(FCM)algorithm; passenger arrival; service time

DOI: 10.3724/SP.J.1249.2020.03314

备注

对平面移动式立体车库客流状态进行聚类有助于获知车库客流信息,指导存取车设备调度优化,提高立体车库的运行效率.提出一种平面移动式立体车库客流状态聚类方法,建立立体车库设备运行时间模型,模拟运行得到不同顾客到达率下的运行参数,对客流模式进行划分并作为初始聚类输入条件,在此基础上使用模糊c均值算法完成对车库客流数据的聚类.以中国西安市某立体车库为研究对象,对其上位机数据库的顾客到达数据进行分析与聚类,定义内部评价指标和相对评价指标,并与其他聚类算法进行对比.实验结果表明,模糊c均值聚类算法实现了立体车库客流状态聚类,且聚类结果可靠、合理地反映出立体车库客流实际情况.

Clustering the flat mobile automated garage passenger flow state can help to obtain the passenger flow information, guide the scheduling optimization and improve the operating efficiency. This paper proposes a method to cluster the passenger flow in flat mobile automated garage. Firstly, we establish the travel time model for the automated garage. After getting simulated operational indicator under different arrival rates, we obtain the passenger flow pattern divisions as clustering input criteria, and use the fuzzy c-means(FCM)algorithm to cluster the garage passenger flow data. Finally, we take an automated garage in Xi'an, China as an example, and obtain the customer arrival data from its upper computer database, define the external evaluation index and relative evaluation index to compare with other clustering algorithms. The experimental results show that the fuzzy c-means clustering algorithm realizes the passenger flow status clustering of stereo garage, and the clustering results reflect the actual passenger flow situation reliably and reasonably.

·