传统PSO算法及其初始化.zip_PSO_初始化

function [x,endPop]=pso(bounds,w,c1,c2,termFns,termOps,evalFN,... evalOps,startPop,pbest,gbest); % PSO run a particle swarm algorithm % [x,endPop,bestPop,trace]=pso(bounds,w,c1,c2,termFns,termOps,evalFn,evalOps, % startPop,pbest,gbest); % % Output Arguments: % x - the best solution found during the course of the run % endPop - the final population % bPop - a trace of the best population % traceInfo - a matrix of best and means of the ga for each generation % % Input Arguments: % bounds - a matrix of upper and lower bounds on the variables % w - the inertia weight % c1,c2 - the acceleration constants % termFns - name of the .m termination function (['maxGenTerm']) % termOps - options string to be passed to the termination function % ([100]). % evalFn - the name of the evaluation .m function % evalOps - options to pass to the evaluation function ([NULL]) % startPop - a matrix of solutions that can be initialized % from initialize.m % pbest - the initial best solution of each particle % gbest - the initial best solution of generation 1 % This program is written by Li Jun on Dec,15,2003.All rights reserved D = size(bounds,1); num = size(startPop,1); gen = 1 ; done = 0 ; vnow = startPop(:,(D+1):2*D); xnow = startPop(:,1:D); xnew = xnow; pb = pbest; gb = gbest; while(~done) %Caculate the fitness value of position xnow xnow=xnew; for m=1:num eval(['[sol,val]=' evalFN '(xnow(m,1:2),1);']); value(m,1)=val; end %replace the pbest for m=1:num if pb(m,D+1) < value(m,1) pb(m,1:D)=xnow(m,1:D); pb(m,D+1)=value(m,1); end end %replace the gbest [maxval,col]=max(value); if gb(D+1) < maxval gb=[xnow(col,:),maxval]; bestg=gen; end %caculate the vnew and xnew vnow = w*vnow+c1*rand*(pb(:,1:D)-xnow)+c2*rand*([gb(1)*ones(num,1),... gb(2)*ones(num,1)]-xnow); xnew = xnow+vnow; if (100-gb(3)<1e-6) | (gen >= termOps) done=1; else gen=gen+1; w=w-0.01; % w declines linearly fprintf(1,'%d ',gen); end end endPop=[xnow,value]; x=gb; fprintf(1,'\n w= %f',w); fprintf(1,'\n bestgen= %d',bestg);