联合FCM与群集蜘蛛优化SVR的短时交通流量预测

Short-term Traffic Flow Prediction Algorithm Based on FCM and Optimized SVR with Social Spider Optimization Algorithm

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%。

The short-term traffic flow forecasting problem of intelligent transportation system（ITS） is studied,and a traffic flow prediction algorithm based on combined FCM and optimized SVR with social spider optimization algorithm is proposed.Firstly,the FCM clustering method is used to process the traffic flow data,and the time sequence data modules based on time node segmentation are obtained,which effectively reduce the error caused by the difference of the data.Then,the SVR prediction model based on social spider optimization（SSO） algorithm is constructed.For the disadvantage of parameter selection of SVR,the social hierarchy is introduced to SSO,and the spider population is divided into three different classes dynamically.According to the individual fitness in different classes,the adaptive competition, ＂fast search＂ and reverse learning mechanism are designed,helping to improve the accuracy of SSO.Finally,the optimized SVR with SSO is used to predict the short-term traffic flow for each time sequence data module.Simulation results show that,Compared with other prediction algorithms,the average absolute deviation of the algorithm is reduced by 38.4%to 53.8%.