基于matlab的无约束多维极值问题优化

function [x,minf] = minSimpSearch(f,X,alpha,sita,gama,beta,var,eps) format long; if nargin == 7 eps = 1.0e-6; end N = size(X); n = N(2); FX = zeros(1,n); while 1 for i=1:n FX(i) = Funval(f,var,X(:,i)); end [XS,IX] = sort(FX); Xsorted = X(:,IX); px = sum(Xsorted(:,1:(n-1)),2)/(n-1); Fpx = Funval(f,var,px); SumF = 0; for i=1:n SumF = SumF + (FX(IX(i)) - Fpx)^2; end SumF = sqrt(SumF/n); if SumF <= eps x = Xsorted(:,1); break; else x2 = px + alpha*(px - Xsorted(:,n)); fx2 = Funval(f,var,x2); if fx2 < XS(1) x3 = px + gama*(x2 - px); fx3 = Funval(f,var,x3); if fx3 < XS(1) Xsorted(:,n) = x3; X = Xsorted; continue; else Xsorted(:,n) = x2; X = Xsorted; continue; end else if fx2 < XS(n-1) Xsorted(:,n) = x2; X = Xsorted; continue; else if fx2 < XS(n) Xsorted(:,n) = x2; end x4 = px + beta*(Xsorted(:,n) - px); fx4 = Funval(f,var,x4); FNnew = Funval(f,var,Xsorted(:,n)); if fx4 < FNnew Xsorted(:,n) = x4; X = Xsorted; continue; else x0 = Xsorted(:,1); for i=1:n Xsorted(:,j) = x0 + sita*(Xsorted(:,j) - x0); end end end end end X = Xsorted; end minf = Funval(f,var,x); format short;