摘要
为提升高速公路交通运行状态评价的效果,提出GA-KFCM(genetic algorithm-kernel fuzzy C-means,基于遗传算法改进的核模糊C均值)聚类算法,并结合实例数据对不同方案的分类效果开展验证分析。首先,分析高速公路交通运行状态评价的范围及等级;然后,提出核函数改进的KFCM(kernel fuzzy C-means,核模糊C均值)聚类算法。在此基础上,采用遗传算法弥补初始化聚类中心随机的缺陷,考虑到在选取不同参数时判别模型的差异较大,结合实例数据对改进前后模型的交通运行状态开展聚类分析,并采用综合指标评估不同试验方案的优劣。试验结果表明:与FCM(fuzzy C-means,模糊C均值)聚类算法相比,GA-KFCM算法的聚类效果提升5倍左右;三维交通参数的交通运行状态判别可靠度最高。
In order to improve the effect of highway traffic flow status evaluation,the study proposes an improved KFCM algorithm based on genetic algorithm,and verifies the classification effect of different schemes combined with case data.Firstly,the evaluation range and grade of highway traffic operation status are analyzed,and then a fuzzy C-means clustering analysis algorithm optimized by kernel function is proposed.On this basis,genetic algorithm is adopted to make up for the defects of initial clustering center randomness.Combined with the case data,cluster analysis is carried out on the traffic running state of the improved model before and after the selection of different parameters,and the advantages and disadvantages of different experimental schemes are evaluated by comprehensive indicators.The experimental results show that the clustering performance of GA-KFCM algorithm is about 5 times higher than that of FCM algorithm.In addition,the reliability of judging traffic operating state under three-dimensional traffic parameters is the highest.
作者
倪琪
牛传同
方为舟
NI Qi;NIU Chuantong;FANG Weizhou(China Design Group Co.,Ltd.,Nanjing 210014,China;Jiangsu Provincial Transportation Engineering Construction Bureau,Nanjing 210004,China;Jiangsu Zhonglu Engineering Technology Research Institute Co.,Ltd.,Nanjing 211899,China)
出处
《现代交通技术》
2024年第2期68-73,共6页
Modern Transportation Technology
关键词
高速公路
交通运行状态评价
聚类分析
核函数
遗传算法
highway
traffic operation status evaluation
cluster analysis
kernel function
genetic algorithm
