基于遗传算法的二元决策树方法在机车当量公里数学模型中的应用研究  

Application of genetic algorithm-based binary decision tree method to mathematical model for equivalent locomotive running kilometers

作　　者：齐金平[1] 宋薇[2] 李宏亮[1] 杨团名[1] 

机构地区：[1]兰州交通大学机电技术研究所,甘肃兰州730030 [2]兰州交通大学自动化与电气工程学院,甘肃兰州730030

出　　处：《内燃机车》2010年第2期13-15,19,共4页Diesel Locomotives

摘　　要：基于遗传算法构造二元决策树的方法,不但可以求解出树结点的线性权值矢量,同时也确定了二元决策树的结构。将该方法应用于机车当量公里数学模型中,提出了一种确定机车当量公里系数的新方法,利用二元决策树分类预测结果,科学、合理地确定不同工况下的机车当量公里数学模型。The genetic algorithm-based binary decision tree method can not only solve a linear decision tree node's weighted vector,but also can determine the binary decision tree's structure.The method is applied to the mathematical model of equivalent locomotive running kilometers,and a nwe method to determine equivalent locomotive running kimeter factor.By using the binary decision tree classification predictions,the mathematical model of equivalent locomotive running kilometers at different operating couditions can be determined reasonably.

关 键 词：遗传算法 决策树 当量公里系数 机车当量公里 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]