基于GA-GM(1,N,α)幂模型的铁路客运量预测

Railway Passenger Volume Forecasting based on GA-GM(1,N,α) Power Model

针对在铁路客运量预测领域传统的灰色预测模型不能反映真实系统的非线性结构特点及其背景值的赋值不合理的问题,提出使用对系统相关因素引入幂指数且经过背景值优化的GM(1,N,)幂模型进行客运量预测。背景值优化时设置待定参数,利用线性组合结构重新计算背景值。对此模型产生的较多的待定参数,采用能够并行运算、全局寻优的遗传算法进行一次性求解。最后使用此模型对甘肃省铁路客运量进行建模预测,并与传统GM(1,N)模型、GM(1,N)幂模型进行对比分析。结果证明,GM(1,N,)幂模型具有更高的预测精度,对铁路客运量预测有一定的应用和研究价值。

In order to reflect the nonlinear structural characteristics of the real system in railway passenger volume forecasting and set the background value of gray prediction models in a reasonable way,a GM（1,N,α） power model with background value optimization is proposed for forecasting passenger volume. The linear combined structure with a probable parameter is used to recalculate the background value and more probable parameters are solved by a genetic algorithm that can perform parallel operation and global optimization for one-time solution. Finally,forecasting railway passenger volume in Gansu Province is conducted with this model. By comparing it with the traditional GM（1,N,α） model and GM（1,N,α） power model,we find that GM（1,N,α） power model has higher prediction accuracy and proves applicable in railway passenger traffic forecasting.