二次插值对称点算法求解一维搜索问题  

Symmetric Point Based Quadratic Interpolation Algorithm for One-dimensional Searching Problem

作　　者：杨爽[1] 贾礼平[1] 

机构地区：[1]乐山师范学院数学与信息科学学院,乐山614000

出　　处：《乐山师范学院学报》2013年第12期3-5,9,共4页Journal of Leshan Normal University

基　　金：四川省教育厅科研项目(No.13ZB0107);乐山师范学院项目(No.JG11-YB11)

摘　　要：插值法是一类用插值多项式来逼近未知或复杂函数的方法。本文基于二次插值,将插值多项式的极小点和其对称点作为搜索区间的两个探索点,通过不断缩小搜索区间,求解一维搜索问题的最优解。本文给出了二次插值对称点的算法,并用0.618法进行了数值比较。结果表明,新算法比0.618法效果好。Interpolation method is a kind of method used interpolation polynomial to approximate unknown or complicated function. Based on quadratic interpolation method, the minimal point and its symmetric point are taken to construct the searching interval. By narrowing the search interval, the optimal solution is obtained for the one-dimensional searching problem. The concrete procedure of the proposed algorithm and numerical results are given. It shows that the new algorithm is more effective than 0.618 method.

关 键 词：一维搜索 优化方法 二次插值 二次插值对称点算法 

分 类 号：O221.2[理学—运筹学与控制论]