中文核心期刊
CSCD来源期刊
中国科技核心期刊
RCCSE中国核心学术期刊
Journal of Chongqing Jiaotong University(Natural Science)

Journal of Chongqing Jiaotong University(Natural Science) ›› 2025, Vol. 44 ›› Issue (12): 53-61.DOI: 10.3969/j.issn.1674-0696.2025.12.07

• Traffic & Transportation + Artificial Intelligence • Previous Articles     Next Articles

Expressway Traffic Flow Data Repair Based on Hybrid Strategy Improved SSA-FCM

HE Qingling1, 2, LIU Jing3, WANG Changfeng4, CHENG Rui2   

  1. (1. School of Traffic and Transportation, Lanzhou Jiaotong University, Lanzhou 730070,Gansu, China; 2. Guangxi Key Laboratory of ITS, Guilin University of Electronic Technology, Guilin 541004, Guangxi, China; 3. College of Civil Engineering and Transportation, Northeast Forestry University, Harbin 150040, Heilongjiang, China; 4. Guangdong Zhongzhi Trace Judicial Identification Institute, Foshan 528011, Guangdong, China)
  • Received:2024-09-09 Revised:2025-04-06 Published:2025-12-25

基于混合策略改进SSA-FCM的快速路交通流数据修复研究

何庆龄1,2，刘静3，王昌锋4，程瑞2   

  1. (1. 兰州交通大学 交通运输学院，甘肃 兰州 730070； 2. 桂林电子科技大学 广西智慧交通重点实验室，广西 桂林 541004； 3. 东北林业大学 土木与交通学院，黑龙江 哈尔滨 150040； 4. 广东中智痕迹司法鉴定所，广东 佛山 528011)
  • 作者简介:何庆龄（1994—），男，甘肃靖远人，讲师，博士，主要从事交通规划、智能优化算法方面的研究。E-mail：qinglinghe@yeah.net 通信作者：程瑞（1992—），男，山东菏泽人，副教授，博士，主要从事交通安全方面的研究。E-mail：ruicheng1992@yeah.net
  • 基金资助:
    广西自然科学基金项目(2023GXNSFAA026359)； 桂林市科技计划项目(20230120-7); 兰州交通大学青年科学基金项目（2025021）； 兰州市哲学社会科学规划项目(25-A21)

PDF

Abstract: In order to address the deficiencies in accuracy and applicability of the clustering and repair results of expressway traffic flow data, an expressway traffic flow data repair model based on hybrid strategy improved SSA-FCM was constructed. Firstly, the SSA algorithm was improved by using hybrid strategies such as Logistic-Tent combination mapping, elite reverse learning and Cauchy mutation hybrid mechanism, so as to improve the diversity and quality of its population and overcome the issues of SSA algorithm being prone to local optima and premature convergence. Secondly, the ambiguity index, clustering center and number of FCM were determined by using hesitant fuzzy theory and ISSA algorithm. Finally, based on the vehicle trajectory data of expressway, the repair effect of the proposed model was compared and analyzed. The research results show that the mean value and standard deviation of the numerical simulation results of the ISSA algorithm in the eight benchmark test functions are closer to the optimal optimization values. The average absolute error (EMA), root mean square error (ERMS), and average absolute percentage error (EMAP) of the data repair results obtained by ISSA-FCM model under different random missing rate conditions are 4.1 km/h, 4.3 km/h and 7.1%, respectively. Compared to SSA-FCM, PSO-FCM, GA-FCM, LSTM, and ARIMA, the above errors are reduced by 13.9% to 58.3%, 40.3% to 68.2%, and 12.4% to 56.6%, respectively.

Key words: traffic engineering; expressway traffic flow; missing data repair; Logistic-Tent combination mapping

摘要： 为解决快速路交通流数据聚类修复结果精度和适用性不足的缺陷，构建了基于混合策略改进SSA-FCM的快速路交通流数据修模型。使用Logistic-Tent组合映射、精英反向学习与柯西变异混合机制等混合策略对SSA算法进行改进，以提升其种群多样性和质量，克服SSA算法易陷入局部最优和过早收敛的问题；采用犹豫模糊理论和ISSA算法，确定FCM的模糊度指数、聚类中心和数目；以快速路车辆轨迹数据为基础，对比分析了该模型的修复效果。研究结果表明：ISSA算法在8个基准测试函数数值仿真结果中的平均值和标准差均更接近最佳优化值，ISSA-FCM模型对随机不同缺失率条件下数据修复结果的平均绝对误差(EMA)、均方根误差(ERMS)、平均绝对百分比误差(EMAP)分别为4.1 km/h、 4.3 km/h、 7.1%；相较于SSA-FCM、 PSO-FCM、 GA-FCM、 LSTM和ARIMA分别降低13.9%~58.3%、 40.3%~68.2%和12.4%~56.6%。

关键词: 交通工程；快速路交通流；缺失数据修复；Logistic-Tent组合映射

CLC Number: