优化方法与实践-进退法（matlab实现）

最新推荐文章于 2023-07-15 21:31:19 发布

风君子诶

最新推荐文章于 2023-07-15 21:31:19 发布

阅读量336

点赞数

分类专栏：优化方法与实践文章标签：算法

本文链接：https://blog.csdn.net/cwzbg/article/details/127039293

版权

优化方法与实践专栏收录该内容

6 篇文章 1 订阅

订阅专栏

作用：是一种确定单变量函数的的搜索区间的数值方法

算法：

代码：

%Author:风君子
%Date:2022.9.25
%Course:Optimization
%Forward-backward method
function[a,b,c] = Forward_backward(phi,alpha0,h0)
%输出：[a,b]为输出搜索区间;c为中间点
%输入：phi，函数；alpha0，起始点；h0起始位移
alpha = alpha0;
alphak = alpha0;
hk = h0;
alphak1 = alphak + hk;
phik = feval(phi,alphak);
phik1 = feval(phi,alphak1);
k=0;
while((phik1<phik)|(k==0))
    if(phik1<phik)
        hk = 2*hk;
        alpha = alphak;
        alphak = alphak1;
        phik = phik1;
        k = k + 1;
        
        alphak1 = alphak + hk;
        phik1 = feval(phi, alphak1);
    else if(k==0)
            hk = - hk;
            alpha = alphak1;
            alphak = alphak;
            phik = phik;
            k=1;
            
            alphak1 = alphak + hk;
            phik1 = feval(phi, alphak1);
        end
    end
end
    a = min(alpha,alphak1);
    b = max(alpha,alphak1);
    c = alphak;