function BCDusingDNN(nsize,tdata,threshold)
dnnobj=dnn(nsize);
dnnobj.initializeweights();
mat=dnnobj.weights;
popsize=100;
solperpop=20;
nogen=100;
vbar_t=1;
universes=convertmat2vect(mat,popsize);
iter=1;maxiter=1000;bestfit=inf;
while bestfit > threshold && iter< maxiter
   [output,universes,fitval]=dnnobj.train('GMVO',tdata,popsize,solperpop,nogen,universes);
   [universes]=sort_universes(universes,fitval);
   vbar_t=updatavbat_t(vbar_t);
   dnnobj.updateweights(universes);
   bestfit=getbestfitval(fitval);
end