对比GA和SA两种优化算法,动态显示优化算法跳出局部最优值

clc; clear; close all; warning off; addpath(genpath(pwd)); f = @objectivefunction; % objective function we want to minimize % f = @objectivefunction2; % objective function we want to minimize a0 = 0; % starting point T = 5; % starting Temperature rho = .95;% temperature change ratio des_space = linspace(-6,6,400); % design space iter = 0; %plot objfunc for visualization purposes fig = line(des_space,f(des_space)); fig.LineWidth = 2; xlab = 'Design Space'; ylab = 'ObjFn Value'; h = gca; h.XLabel = xlabel(xlab); h.YLabel = ylabel(ylab); C0 = f(a0);fig2 = line(a0,C0);fig2.Marker = 'o'; h = gca; txt = sprintf('Temperature = % 2.2f', T); xl = mean(des_space); yl = f(xl); t = text(xl, yl,txt); mins = []; minserror = 10; fncalls = 0; while T >= .01 && minserror > 1e-2 while iter <= 20 && minserror > 1e-2 delete(t); a = datasample(des_space,1); C = f(a); if (C < C0) || (exp((C0-C)/T) > rand) a0 = a; C0 = C; end mins = [mins C0]; if length(mins) > 10 minserror = norm(mins - mean(mins)); mins = []; end iter = iter + 1; delete(fig2); fig2 = line(a0,C0); fig2.Marker = 'o'; fig2.MarkerSize = 10; fig2.MarkerFaceColor = 'green'; fig2.MarkerEdgeColor = 'black'; titl = sprintf('Design point = %1.2f, Objfn value = %2.2f', a0, C0); txt = sprintf('Temperature = % 2.2f', T); t = text(xl,yl,txt); set(t, 'String',txt); % fig2.Annotation = txt; h.Title = title(titl); h.TitleFontWeight = 'normal'; h.XLabel = xlabel(xlab); h.YLabel = ylabel(ylab); drawnow; pause(0.001); end fncalls = fncalls + iter; iter = 1; T = rho*T; end disp('function calls'); disp(fncalls); disp('minimum location'); disp(a0); disp('minimum'); disp(C0);