用遗传算法改进粒子群GA-PSO算法.rar

%% Start of the program. clc clear close all %% Problem Statement (Traveling Sealsman Problem) NPar = 8; % Number of Optimization Parameters VarLow =[0 0 0 0.03 0.01 -0.05 -0.06 -0.06]; % Lower Limmit of problem VarHigh =[1 1 0.12 0.15 0.11 0.05 0.04 0.04]; % Upper Limmit of problem FunName = 'Myfit'; %% Algoritm Parameters PopSize = 30; ChromLenght = NPar; MaxIterations = 50; KeepPercent = 50/100; CrossPercent = 70/100; MutatPercent = 1 - KeepPercent - CrossPercent; SelectionMode = 1; KeepNum = round(KeepPercent * PopSize); CrossNum = round(CrossPercent * PopSize); if mod(CrossNum,2) ~= 0 CrossNum = CrossNum + 1; end MutatNum = PopSize - KeepNum - CrossNum; %% Initial Population AA=rand(PopSize,NPar); AB=VarHigh - VarLow; ABC=[]; for i=1:8,C=AA(:,i)*AB(:,i)+VarLow(1,i),ABC=[ABC C],end; Pop = ABC clear AA AB ABC C; Cost = feval(FunName,Pop); [Cost Indx] = sort(Cost); Pop = Pop(Indx,:); %% MAIN LOOP MinMat = []; MeanMat = []; for Iter = 1:MaxIterations %% Select Keep Pop(1:KeepNum,:) = PSO_Fcn(Pop(1:KeepNum,:)); %% CrossOver SlectedIndexes = SelectParents_Fcn(Cost,CrossNum,SelectionMode); CrossPop = []; for ii = 1:2:CrossNum Par1Index = SlectedIndexes(ii); Par2Index = SlectedIndexes(ii+1); Par1 = Pop(Par1Index,:); Par2 = Pop(Par2Index,:); [Off1 , Off2] = CrossOver_fcn(Par1,Par2); CrossPop = [CrossPop ; Off1 ; Off2]; end Pop(KeepNum+1:KeepNum+CrossNum,:) = CrossPop; %% Mutation AA=rand(MutatNum,NPar) AB=VarHigh - VarLow; ABC=[]; for i=1:8,C=AA(:,i)*AB(:,i)+VarLow(1,i),ABC=[ABC C],end; Pop(KeepNum+CrossNum+1 : end , :)= ABC; clear AA AB ABC; [m,n] = size(Pop) %% NewPopulation for i=1:n jj=1; while jj<=m if Pop(jj,i)>VarHigh(1,i),Pop(jj,i)=VarHigh(1,i),end; if Pop(jj,i)<VarLow(1,i),Pop(jj,i)=VarLow(1,i),end; jj=jj+1; end; end; Cost = feval(FunName,Pop); [Cost Indx] = sort(Cost); Pop = Pop(Indx,:); MinMat = [MinMat min(Cost)]; MeanMat = [MeanMat mean(Cost)]; %% Display plot(MinMat,'r','Linewidth',2.5); xlabel('Iterations') ylabel('Cost') hold on plot(MeanMat,'b:','Linewidth',2); hold off xlim([1 MaxIterations]) pause(0.05) end %% Final Result Display BestSolution = Pop(1,:) BestCost = Cost(1) fval=BestCost;