利用MATLAB求解无约束优化问题.ppt

命令格式为:[x，fval，exitflag，output]=fminunc(fun,x0，options)；或[x，fval，exitflag，output]=fminsearch(fun,x0,options)；[3]fminunc为中型优化算法的步长一维搜索提供了两种算法，由options中参数LineSearchType控制：LineSearchType=’quadcubic’(缺省值)，混合的二次和三次多项式插值；LineSearchType=’cubicpoly’，三次多项式插值例1minf(x)=(4x12+2x22+4x1x2+2x2+1)*ex14用fminsearch函数求解4.用fminunc函数options11=optimset('HessUpdate','dfp')[x11,fval11,exitflag11,output11]=fminunc('fun2',[-1.22],options11)pauseoptions12=optimset('HessUpdate','dfp','LineSearchType','cubicpoly')[x12,fval12,exitflag12,output12]=fminunc('fun2',[-1.22],options12)pauseoptions21=optimset('HessUpdate','bfgs')[x21,fval21,exitflag21,output21]=fminunc('fun2',[-1.22],options21)pauseoptions22=optimset('HessUpdate','bfgs','LineSearchType','cubicpoly')[x22,fval22,exitflag22,output22]=fminunc('fun2',[-1.22],options22)pauseoptions31=optimset('HessUpdate','steepdesc')[x31,fval31,exitflag31,output31]=fminunc('fun2',[-1.22],options31)pauseoptions32=optimset(,'HessUpdate','steepdesc','MaxIter',8000,'MaxFunEvals',8000)[x32,fval32,exitflag32,output32]=fminunc('fun2',[-1.22],options32)pauseoptions33=optimset('HessUpdate','steepdesc','MaxIter',9000,'MaxFunEvals',9000)[x33,fval33,exitflag33,output33]=fminunc('fun2',[-1.22],options33)（ii）编写M文件zuisu.mx=[2;2];[f0,g]=detaf(x);whilenorm(g)>0.001d=-g'/norm(g);t=1.0;f=detaf(x+t*d);whilef>f0t=t/2;f=detaf(x+t*d);endx=x+t*d[f0,g]=detaf(x)end