摘要
本文将函数负梯度的概念应用于一维搜索中,提出了一维搜索的新方法——双切线法和双割线法.特别是双割线法与0.618法和抛物线插值法相比.可使一维搜索的效率明显提高.
By the use of the concept of functional negative gradient, this paper states new algorithms or the unidimensional search. the double-cuttingLine(DCL) and the double-secant Line(DSL) methods.Especially, as com-pared with the 0.618 method and the parabolic function-fitting method, the doub-le-secant line method makes the efficiency of the unidimensional search raise obvi-ously.
关键词
一维搜索
双切线法
双割线法
Optimum search
Gradient algorithm
Steepest descent method
