论文部分内容阅读
针对高速公路短时交通量实时性、波动性和非线性的特点,将参数投影寻踪回归(parameter projection pursuit regression,PPPR)方法应用于高速公路短时交通量预测.采用可变阶的正交Hermite多项式拟合其中的岭函数,运用最小二乘法确定多项式权系数c.为了更好地优选PPPR模型的投影方向a和岭函数个数M,利用混沌云粒子群算法对模型参数进行优选.提出了在外层优化岭函数个数M的同时,利用CCPSO算法在内层优化最佳投影方向a的CCPSO-PPPR混合优化高速公路短时交通量预测模型.将路段前几个时段交通量、天气因素和出行日期作为影响因素输入.实例预测与模型对比结果表明,该模型取得了更好的预测效果,绝对误差控制在[-6,6]以内,可有效应用于高速公路短时交通量预测.
According to the characteristics of real-time, volatility and nonlinearity of expressway short-term traffic, parameter projection pursuit regression (PPPR) method is applied to expressway short-term traffic forecasting. Hermite polynomial is used to fit the ridge function and the least square method is used to determine the polynomial weight coefficient c. In order to optimize the projection direction a of the PPPR model and the number of ridge function M, the chaos cloud particle swarm algorithm is used to optimize the model parameters. In order to optimize the short-term traffic volume of expressway with CCPSO algorithm, CCPSO-PPPR hybrid optimizes the short-term traffic volume forecasting model of expressway with CCPSO algorithm in the inner layer.Considering the traffic volume, weather factors And travel date as input factors.Compared with the model predictions, the results show that the model has better forecasting results and the absolute error is controlled within [-6,6], which can be effectively applied to expressway short-term traffic volume forecasting.