对于非负权重最优组合预测模型,文章提出一种新的求解方法—Wolfe方法。它以Kuhn-Tucker条件为基础把二次规划问题转化成线性规划问题。文章给出具体的实现方法并总结出解决这类问题通用的线性规划模型。最后,通过实例说明这种方法是行之有效的。
The paper put forward a new method-Wolfe's method to solve the optimal combined torecasts of non-negative weights. It transformed quadratic programming problem to linear programming problem based on Kuhn-Tucker condition, gave the detailed procedure, and generalized the linear programming model which could solve the problems. Finally, the method was proved to be valid through a example.