matlab-基于CSO猫群优化算法的多目标优化matlab仿真-源码

clc; clear; close all; warning off; addpath(genpath(pwd)); % Problem Definition nVar=[3 0]; % Number of Decision Variables VarSize=[1 , nVar(1) ; 1 , nVar(2)]; % Size of Decision Variables Matrix VarMin=[-5 , 0]; % Lower Bound of Variables VarMax=[5 , 0]; % Upper Bound of Variables % CSO Parameters MaxIt=100; % Maximum Number of Iterations nPop=50; % Population Size (Swarm Size) nTest = 0.1 * nPop; w=1; % Inertia Weight wdamp=0.99; % Inertia Weight Damping Ratio c1=2; % Global Learning Coefficient MR = 0.9; % 1 = all cats in seeking and 0 = all cats in tracing nGrid=5; % Number of Grids per Dimension alpha=0.1; % Inflation Rate beta=2; % Leader Selection Pressure reInitPer = 0.2; % Velocity Limits VelMax=0.1*(VarMax-VarMin); VelMin=-VelMax; % Initialization empty_particle.Position=[]; empty_particle.Cost=[]; empty_particle.Velocity=[]; empty_particle.flag = false; empty_particle.Best.Position=[]; empty_particle.Best.Cost=[]; empty_particle.Rank=[]; empty_particle.DominationSet=[]; empty_particle.DominatedCount=[]; empty_particle.CrowdingDistance=[]; empty_particle.GridIndex=[]; empty_particle.GridSubIndex=[]; empty_particle.index = 0; particles=repmat(empty_particle,nPop,1); TestParticles = repmat(empty_particle , nTest , 1); % Test Swarm % Initialize Population for i=1:nPop % Initialize Position particles(i).Position(1:nVar(1))=unifrnd(VarMin(1),VarMax(1),VarSize(1,:)); particles(i).Position(nVar(1)+1:nVar(2)+1)=unifrnd(VarMin(2),VarMax(2),VarSize(2,:)); % Initialize Velocity particles(i).Velocity(1:nVar(1))=zeros(VarSize(1,:)); particles(i).Velocity(nVar(1)+1:nVar(2)+1)=zeros(VarSize(2,:)); % Evaluation particles(i).Cost=Benchmark(particles(i).Position , 0); particles(i).index = i; % Update Personal Best particles(i).Best.Position=particles(i).Position; particles(i).Best.Cost=particles(i).Cost; end % Initialize Test Population for i=1:nTest % Initialize Position TestParticles(i).Position(1:nVar(1))=unifrnd(VarMin(1),VarMax(1),VarSize(1,:)); TestParticles(i).Position(nVar(1)+1:nVar(2)+1)=unifrnd(VarMin(2),VarMax(2),VarSize(2,:)); % Evaluation TestParticles(i).Cost=Benchmark(TestParticles(i).Position , 0); end % BestCost=zeros(MaxIt,1); % Non-Dominated Sorting [particles , F]=NonDominatedSorting(particles); % Calculate Crowding Distance particles=CalcCrowdingDistance(particles,F); % Sort Population [particles , F]=SortPopulation(particles); for j = 1 : length(particles) particles(j).index = j; end % Determine Domination F1 = particles(F{1}); Grid=CreateGrid(F1,nGrid,alpha); for i=1:numel(F1) F1(i)=FindGridIndex(F1(i),Grid); end % CSO Movement for it=1:MaxIt TestBackupCosts = [TestParticles.Cost]; TestNewCosts = zeros(2 , nTest); for i=1:nTest TestParticles(i).Cost = Benchmark(TestParticles(i).Position , it); TestNewCosts(: , i) = TestParticles(i).Cost; end tf = isequal(TestBackupCosts , TestNewCosts); if(tf == 0) OldIndex = [particles.index]; OldIndex = OldIndex'; for i=1:nPop particles(i).Cost = Benchmark(particles(i).Position , it); end [particles , F]=NonDominatedSorting(particles); particles = CalcCrowdingDistance(particles,F); [particles , F]=SortPopulation(particles); NewIndex = [particles.index]; NewIndex = NewIndex'; OptIndex = BordaCountAlgorithm([OldIndex , NewIndex]); NumOfNoChange = floor((1 - reInitPer) * length(OptIndex)); for i = NumOfNoChange : length(OptIndex) particles(OptIndex(i)).Position(1:nVar(1))=unifrnd(VarMin(1),VarMax(1),VarSize(1,:)); particles(OptIndex(i)).Position(nVar(1)+1:nVar(2)+1)=unifrnd(VarMin(2),VarMax(2),VarSize(2,:)); % Evaluation particles(OptIndex(i)).Cost=Benchmark(particles((OptIndex(i))).Position , it); end [particles , F]=NonDominatedSorting(particles); particles = CalcCrowdingDistance(particles,F); [particles , F]=SortPopulation(particles); for j = 1 : length(particles) particles(j).index = j; end F1 = particles(F{1}); Grid=CreateGrid(F1,nGrid,alpha); for i=1:numel(F1) F1(i)=FindGridIndex(F1(i),Grid); end end indices = randperm(nPop , floor(nPop*MR)); SeekingPop = []; TracingPop = []; leader=SelectLeader(F1,beta); for i=1:nPop ind = find(indices == i); if(~isempty(ind)) % Seeking Mode [SeekingParticles, leader] = Seeking( particles(i) , leader , VelMax , VelMin , VarMax , VarMin , VarSize , nVar , it); SeekingPop = [SeekingPop , SeekingParticles]; else % Tracing Mode [ particles(i) , leader ] = Tracing( particles(i) , leader , VelMax , VelMin , VarMax , VarMin , VarSize , nVar , w , c1 , it); TracingPop = [TracingPop , particles(i)]; end end particles = [SeekingPop , TracingPop]; % Non-Dominated Sorting [particles , F]=NonDominatedSorting(particles); % Calculate Crowding Distance particles=CalcCrowdingDistance(particles,F); % Sort Population [particles , F]=SortPopulation(particles); % Truncate particles=particles(1:nPop); % Non-Dominated Sorting [particles , F]=NonDominatedSorting(particles); % Calculate Crowding Distance particles=CalcCrowdingDistance(particles,F); % Sort Population [particles , F]=SortPopulation(particles); for j = 1 : length(particles) particles(j).index = j; end % Store F1 F1=particles(F{1}); Grid=CreateGrid(F1,nGrid,alpha); for i=1:numel(F1) F1(i)=FindGridIndex(F1(i),Grid); end % Show Iteration Information disp(['Iteration ' num2str(it) ': Number of F1 Members = ' num2str(numel(F1))]); % Plot F1 Costs figure(1); PlotCosts(F1); w=w*wdamp; end