基于聚类分析与0-1规划轻型汽车行驶工况构建方法

Method for Construct Driving Condition of Light Vehicles Based on Cluster Analysis and 0-1 Programming

针对汽车行驶工况构建的问题,文中对福州市实际汽车行驶数据进行数据挖掘,提出一种结合聚类分析和“0-1”规划的汽车行驶工况的构建方法.其步骤为:对原始数据预处理,并将行驶数据划分为共1751条运动学片段;对所有片段主成分分析,以其中前两个主成分为特征作K均值聚类;选取具有代表性的片段为备选并作“0-1”规划,得到最优行驶工况构建方案.实验结果表明:该方案与实际数据相比,平均误差比约为4.5%;与国际五项典型工况相比,该方案误差仅为其他方案误差的约10%,证明了构建新型的、符合所在城市的车辆行驶工况的必要性.

Aiming at the problem of building automobile driving conditions,this paper carried out data mining on the actual automobile driving data in Fuzhou,and put forward a method of building automobile driving conditions combining cluster analysis and“0-1”planning.Firstly,the original data was preprocessed,and the driving data was divided into 1751 kinematic segments.Then,the principal components of all fragments were analyzed,and K-means clustering was performed with the first two principal components as characteristics.Selecting representative segments as alternatives and making“0-1”planning,the construction scheme of optimal driving conditions was obtained.Experimental results show that the average error ratio is about 4.5%compared with the actual data.Compared with five typical international operating conditions,the error of this scheme is only about 10% of that of other schemes.