基于云遗传算法优选的SVR交通量预测模型

Urban Traffic Flow Forecasting Base on CCLGA Algorithm Optimizes Parameters of SVR

针对城市主干道交通流量的实时变化和波动性特点,利用支持向量回归机(support vectorregression,SVR)进行城市主干道短时交通流量预测.为了优选SVR模型参数,基于混沌logistic映射和云自适应机制对标准遗传算法进行改进,建立了基于混沌云自适应遗传算法(chaos clouds adaptive genetic algorithm,CCLGA)进行SVR参数优选的CCLGA-SVR城市主干道短时交通流量预测模型.综合考虑了短时交通量各个影响因素,结合实测数据进行了实证预测分析,仿真结果表明文中提出的预测模型精度较高,寻优速度较快,可有效应用于城市主干道短时交通流量预测.

To deal with the real-time changes and volatility of city main road traffic flow, support vec- tor regression （SVR） is applied. To choose the appropriate parameters with best effect, this paper proposes an algorithm named CCLGA algorithm, in which logistic chaos variables and cloud-based a-daptive and genet CCLGA-SVR city model was made, than the existing flow. ic algorithm was integrated. By combining CCLGA algor main road short-time traffic flow prediction model. Finall contrastive analysis represents that this model has faster Slc models, and can be effectively applied to predict city main ithm with SVR we get y the verification of the eed and higher precision road short-term traffic