最优化方法的实验报告程序代码

一维搜索进退法

 

function f=getvalue(t)

f=t^3-2*t+1;

 

function [a,b]=myfun(t,h,alpha)

t0=t;

k=0;

fi0=getvalue(t0);

for count=1:1:100

    t1=t0+h;

    fi1=getvalue(t1);

    if(fi1<fi0)

        h=alpha*h;

        t=t0;

        t0=t1;

        fi0=fi1;

        k=k+1;

    else

        if(k~=0)

            if(t>t1)

                a=t1;

                b=t;

            else

                a=t;

                b=t1;

            end

            break;

        else

            h=-h;

            t=t1;

        end

    end

end

 

结果：

>> [a,b]=myfun(0,1,2)

a=0,b=3

 

 

黄金分割法

function [value,a,b]=myfun(m,n,alpha)

tao=0.618;

ephsilo=alpha;

a=m;

k=0;

b=n;

lamenda=a+(1-tao)*(b-a);

miu=a+tao*(b-a);