基于混沌自适应遗传PPPR城市交通量预测

Prediction of Traffic-Flow Based on Chaos Adaptive Genetic Algorithm and PPPR Model

目的为了提高城市交通量预测精度.方法将参数投影寻踪回归(Parameter Projection Pursuit regression,PPPR)思想应用于城市交通量预测,利用最小二乘法确定Hermit多项式权系数c,在优化岭函数个数M的同时基于混沌自适应遗传算法(Chaos Adaptive Genetic Algorithm,CAGA)内嵌式优化PPPR模型的最佳投影方向a,提出了基于CAGA-PPPR模型进行交通量预测的新方法.结果仿真实验表明该模型的预测效果优于对比模型,仿真预测平均绝对相对误差控制在3.47%以内,提高了城市交通量的预测精度.结论应用于城市交通量的预测具有一定的可行性和实用性.

In order to improve the traffic-flow prediction accuracy,a parameter projection pursuit regression(PPPR)was used in this paper to forecast the traffic-flow of city and increase the forecasting accuracy of PPPR.The ridge functions were analyzed by using the Hermite orthogonal polynomial,and the best projection direction a of the PPPR model was optimized by means of Chaos Adaptive Genetic Algorithm(CAGA)at the same time of optimizing the number of ridge functions M.As a result,a new prediction method for fore...