联合FCM与群集蜘蛛优化SVR的短时交通流量预测

发布时间：2018-04-29 20:23

本文选题：交通流量预测 + 模糊C-均值聚类　；参考：《中国电子科学研究院学报》2017年01期

【摘要】：对智能交通系统(ITS)短时交通流量预测问题进行研究,提出了一种联合FCM与群集蜘蛛优化SVR交通流量预测算法。首先采用FCM聚类方法对交通流量数据预处理,得到基于时间节点分割的时序数据模块,有效降低了数据差异性带来的误差影响;然后构建基于群集蜘蛛优化SVR模型,针对SVR参数选择难题,在群集蜘蛛优化算法中引入社会等级制度,动态的将蜘蛛种群划分为上中下三个阶层,并根据不同阶层个体适应度大小,分别设计自适应竞争、"快搜"以及逆向学习机制,提高了算法寻优精度;最后,运用群集蜘蛛优化SVR对各个交通流量数据时序模块进行预测评估。仿真结果表明,同其它预测算法相比,该算法预测平均绝对误差降低了38.4~53.8%。
[Abstract]:In this paper, the problem of short time traffic flow prediction in intelligent transportation system (its) is studied, and a combined FCM and cluster spider optimization SVR traffic flow prediction algorithm is proposed. Firstly, the traffic flow data is preprocessed by using FCM clustering method, and the time series data module based on time node segmentation is obtained, which effectively reduces the error effect brought by the data difference. Then the optimized SVR model based on cluster spider is constructed. In order to solve the problem of SVR parameter selection, the social hierarchy system is introduced into the swarm spider optimization algorithm, and the spider population is dynamically divided into upper, middle and lower classes, and according to the individual fitness of different classes. Adaptive competition, "quick search" and reverse learning mechanism are designed respectively to improve the accuracy of the algorithm. Finally, the traffic flow data timing modules are predicted and evaluated by cluster spider optimization SVR. The simulation results show that, compared with other prediction algorithms, the prediction average absolute error of this algorithm is reduced by 38.4% or 53.8%.
【作者单位】：广东交通职业技术学院智能交通工程技术应用中心;华南理工大学;
【基金】：国家星火计划项目:基于物联网的光伏大棚智能控制技术应用与示范(2015GA780024) 广东省高等学校优秀青年教师培养计划项目:城市交通信号控制实时评价模型及其优化方法研究(Yq2013180) 广东省高等职业教育品牌专业建设项目:智能交通技术运用(2016gzpp044)
【分类号】：U491.14;TP18

【相似文献】