基于随机灰色蚁群神经网络的集装箱结点站运量预测  

Combined forecasting model and application based on random gray gnt colony neural network

作　　者：朱昌锋[1] 

机构地区：[1]兰州交通大学交通运输学院,甘肃兰州730070

出　　处：《铁道科学与工程学报》2011年第2期81-85,共5页Journal of Railway Science and Engineering

基　　金：教育部"春晖计划"项目(Z2005-1-62008);甘肃省自然科学基金资助项目(ZS022-A25-036)

摘　　要：根据铁路集装箱运量预测受到多因素影响以及非线性的特点,利用随机灰色变量描述预测系统的不确定性,建立了随机灰色预测模型以及基于蚁群算法的递归神经网络模型,在此基础上,提出了一种基于随机灰色蚁群神经网络的集装箱结点站运量预测方法。最后,以兰州铁路局兰州北站为例,对模型的预测精度和有效性进行分析。研究结果表明:基于蚁群算法的递归神经网络模型的预测精度不但大于其他单一预测模型的预测精度,而且明显比其他传统组合预测模型的预测精度,能很好地反映事物发展的规律。According to characteristic of multiple factor and nolinear on freight volume forecasting of railway con- tainer node station, using the uncertainty of forecasting system while representing of stochastic gray variable, models of random gray and recurrent neural network based on ant colony optimization algoritl^lm were proposed. On that basis, a model was built to forecast volume of railway container node station based on random gray ant colony neural network. Finally a case study was carried out to forecast north station of Lanzhou railroad bureau, for analyse forecasting accuracy and validity of this model. The results show that random gray and recurrent neu- ral network based on ant colony optimization algorithm is not only greater than other single forecasting model, but also superior to other combined forecasting model, this model can reflect the law of development of object well and can improve forecast accuracy effectively.

关 键 词：灰色模型 神经网络 随机 蚁群优化算法 

分 类 号：U291.1[交通运输工程—交通运输规划与管理]