为提高整车运输中的车辆装载率、提升车货匹配效率，基于CBR系统思维，提出了一种基于视图相似度的车货匹配方法。首先，通过知识表达系统表征货物和车辆信息，并基于车辆的车型属性和货物的名称属性实现二者的初步分类和匹配；而后，对车辆数据集进行K-Means聚类，并基于马氏距离计算并确定与待匹配货物最近的聚类，实现对视图匹配空间的横向压缩；最后，融合改进传统视图计算方法，并利用欧氏距离计算待匹配货物与既定聚类内各车辆的视图相似度，提交最小视图相似度对应之车辆。通过爬取运满满平台数据进行实验分析，证明所提方法能够显著提升车货匹配装载率，且匹配效率提高了76%左右。

In order to improve the vehicle loading rate and make full use of resources in vehicle transportation, this paper proposes a vehicle-cargo matching method based on view similarity in the way of the CBR system thinking. Firstly, the information of goods and vehicles is represented by the knowledge expression system, and the preliminary classification and matching of the two are realized based on the CR attribute of the vehicle and the N attribute of the goods; then, K-Means clustering is performed on the vehicle data set, and the Mahalanobis distance is used to calculate and match the information. Determine the cluster closest to the goods to be matched, and realize the lateral compression of the view matching space; finally, integrate and improve the traditional view calculation method, and use the Euclidean distance to calculate the view similarity between the goods to be matched and each vehicle in the given cluster, and submit the vehicle corresponding to the minimum view similarity. By crawling the Yunmanman platform data for experimental analysis, it is proved that the proposed method can significantly improve the loading rate of truck-cargo matching, and the matching efficiency is increased by about 76%.

国家社科基金项目“隐性知识深度服务体系研究（19BTQ035）”