基于粒子群算法实现 WSN 网络中传输功率优化附matlab代码zip

%################################################################################################################################################################################# %This MATLAB code was used in paper "Particle swarm optimization implementation for minimal transmission power providing a fully-connected cluster for the Internet of Things" %It was based on this following code: http://www.mathworks.com/matlabcentral/fileexchange/20205-particle-swarm-optimization/content/PSO_Wael/PSO/PSO.m. %Access to paper: http://ieeexplore.ieee.org/abstract/document/7224573/ %This code implements a PSO algorithm to optimize transmission powers for each node in an WSN networks %Authors: Gabriel Lobão, Felipe Reis, Jonathan de Carvalho and Lucas Mendes %################################################################################################################################################################################# clear all; close all for cen=1:10 cenario = strcat('space',int2str(cen)); load(cenario) nomeArquivo = strcat(cenario,'.t2'); fid = fopen(nomeArquivo,'wt'); fprintf(fid,'##################Início do Treinamento##################\n\n'); fprintf(fid,'Cenario = %s\n\n',cenario); for t=1:5 tic fprintf(fid,'##########################################################################\n'); fprintf(fid,' TESTE = %d \n',t); fprintf(fid,'##########################################################################\n'); n = 30; % tamanho do enxame num_passos = 500; % máximo número de passos dim = size(X,2); % dimensão da partícula wmax=0.9; %fator de inércia inicial wmin=0.4; %fator de inércia final c1=2; %coeficiente de aceleração c2=2; %coeficiente de aceleração %Define o limite do espaço de busca pMin = -30; pMax = 0; %Define o limite da velocidade vMax = 5; vMin = -vMax; fprintf(fid,'Configuracoes do PSO\n'); fprintf(fid,'Numero de particulas = %f\n',n); fprintf(fid,'Numero maximo de passos = %f\n',num_passos); fprintf(fid,'Dimensao do problema = %f\n',dim); fprintf(fid,'Coeficiente C1 = %f\n',c1); fprintf(fid,'Coeficiente C2 = %f\n',c2); fprintf(fid,'Inercia (wmax) = %f / (wmin) = %f \n\n',wmax,wmin); fprintf(fid,'pMax = %f / pMin = %f\n',pMax,pMin); fprintf(fid,'vMax = %f / vMin = %f\n',vMax,vMin); fitness=0*ones(n,num_passos); current_fitness =0*ones(n,1); %Inicialização das posições das partículas vetor_potencia = pMin + (pMax - pMin)*rand(dim,n); %Inicialização da velocidade da partícula velocidade = rand(dim,n); %velocidade = vMin + (vMax - vMin)*rand(dim,n); %Calcula fitness inicial das partículas for i=1:n current_fitness(i) = connect_report(X,vetor_potencia(:,i)) ; end %Inicializa o PBEST pbest_fitness = current_fitness ; pbest_position = vetor_potencia ; %Inicializa o GBEST [gbest_fitness,g] = min(pbest_fitness) ; for i=1:n gbest_position(:,i) = pbest_position(:,g) ; end for iter=1:num_passos %Decrementa linearmente o fator de inercia w = wmax - ((wmax-wmin)/num_passos)*iter; %Calcula nova velocidade da particula for i=1:n velocidade(:,i) = w*velocidade(:,i) + c1*(rand*(pbest_position(:,i)-vetor_potencia(:,i))) + c2*(rand*(gbest_position(:,i)-vetor_potencia(:,i))); end %Limita a velocidade da particula for k=1:n index = velocidade(:,k)>vMax; velocidade(index,k)=vMax; index = velocidade(:,k)<vMin; velocidade(index,k)=vMin; end %Calcula nova posicação da particula vetor_potencia = vetor_potencia + velocidade; %Limita o espaço de busca da particula for k=1:n index = vetor_potencia(:,k)>pMax; vetor_potencia(index,k)=pMax; index = vetor_potencia(:,k)<pMin; vetor_potencia(index,k)=pMin; end %Avalia novo fitness da partícula for i=1:n, current_fitness(i) = connect_report(X,vetor_potencia(:,i)) ; end %Calcula pbest for i=1:n if current_fitness(i) < pbest_fitness(i) pbest_fitness(i) = current_fitness(i); pbest_position(:,i) = vetor_potencia(:,i); end end [current_gbest_fitness,g] = min(pbest_fitness); %Calcula gbest if current_gbest_fitness < gbest_fitness gbest_fitness = current_gbest_fitness; for i=1:n gbest_position(:,i) = pbest_position(:,g); end end if(iter==1 || rem(iter,50)==0) sprintf('Iteracao - %4d | Soma = %f\n',iter,gbest_fitness) fprintf(fid,'Iteracao - %4d | Soma = %f\n',iter,gbest_fitness); end end %fim do algoritmo gbest_fitness; d=gbest_position(:,1); fprintf(fid,'\n\n'); fprintf(fid,'Potência de cada nó\n\n'); for i=1:dim fprintf(fid,'N?%i= %f ',i, d(i)); fprintf(fid,'\n'); end fprintf(fid,'\n\n'); tempoGasto = toc; fprintf(fid,'Tempo gasto: %d\n\n',tempoGasto); end %fim for t=1:10 fclose(fid); end