基于规划求解的组合预测模型在道路客运量预测中的应用

发布时间：2019-03-09 14:02

【摘要】：为更好地对道路客运量未来发展趋势进行预测,建立更加精确的预测模型,在分析主要预测方法的基础上,提出一种基于规划求解的组合预测模型。基于灰色模型、一元回归和指数平滑三种预测方法,建立了以历年预测值加权之和与实际值的差值的绝对值之和最小为目标函数,以权重系数非负和权重系数之和为1的规划求解模型。在权重计算过程中,目标函数值随着迭代次数的增加逐渐减少,通过对目标函数值的变化值进行观测,当目标函数的变化值出现拐点时,定义了组合权重迭代次数的确定方法。选取更具有代表意义的"百城百站"旅客发送量为基础数据,建立了道路旅客运输客运量组合预测模型。在规划求解过程中利用Excel中的"规划求解"功能。试验结果表明:随着迭代次数的增加,目标函数变化值逐渐减小,当迭代次数为7时,目标函数变化值出现拐点,以此确定了迭代次数;传统3种预测方法的绝对误差分别是1.26,0.48和2.98,基于规划求解的组合预测模型的绝对误差是0.12,预测精度更高、误差更小,并且该组合预测模型操作方便,可减少单个模型预测的不确定性,可根据上述模型对未来道路旅客发送量进行预测。
[Abstract]:In order to forecast the future development trend of road passenger traffic, a more accurate forecasting model is established. Based on the analysis of the main forecasting methods, a combined forecasting model based on programming solution is proposed. Based on the grey model, univariate regression and exponential smoothing three forecasting methods, the objective function is established, which is the sum of the absolute value of the difference between the weighted sum of the predicted value over the years and the difference between the actual value and the sum of the absolute value as the objective function. The programming solution model is based on the sum of the non-negative weight coefficient and the weight coefficient 1. In the process of weight calculation, the value of objective function decreases gradually with the increase of iteration times. By observing the change value of objective function value, when the change value of objective function appears inflection point, the method of determining the iteration number of combined weight is defined. Based on the data of "Baicheng and 100 stations" which is more representative, a combined forecast model of passenger volume for road passenger transport is established. In the process of planning and solving, the function of "programming solving" in Excel is used. The experimental results show that with the increase of the number of iterations, the change value of the objective function decreases gradually. When the number of iterations is 7, the inflection point of the change value of the objective function appears, and the number of iterations is determined. The absolute errors of the three traditional prediction methods are 1.26,0.48 and 2.98, respectively. The absolute error of the combined forecasting model based on programming solution is 0.12. the prediction precision is higher and the error is smaller, and the combined forecasting model is easy to operate. The uncertainty of single model prediction can be reduced, and the future road passenger traffic can be predicted according to the above model.
【作者单位】：交通运输部科学研究院;
【基金】：交通运输部建设科技项目(2015 318 J36 110) 交通运输战略规划政策研究项目(2016-1-4) 中央级公益性科研院所基本科研业务费项目(2016 6110)
【分类号】：U492.413

【相似文献】