making function of a fixed point algorithm

I have the following fixed point algorithm (in R)

delta_<-matrix(1,100,1)
f_<-matrix(0,1000,100)

for (iterate in 1:1000){
  g0<-1

  for (i in 1:1000){
    f0<-t(exp(delta_+ g*v_xj[i,1:100])/(1+sum(exp(delta_+g*v_xj[i,1:100]))))
    f_[i,1:100]<-f0
  }
  s_<-t(matrix(1,1,1000)%*%f_/1000)
  s_0<-1-sum(s_)
  lns_<-log(s_/s_0)
  delta_<-delta_+lnsj-lns_
  if (abs(lnsj[1]-lns_[1])<=aprox)
    break
}

but i want to make them a function in g... i need to obtain many delta for many g ...so i think than the most efficient way is make the algoritm be a function in g.. but when a try

fixed<- function(g){ algoritm.....
result<-list(delta,lns_)
retu(result)
}

the parameters not satifishes the condition

abs(lnsj[1]-lns_[1])<=aprox

in other words, when i a make the function the algorithm doesnt work... someone can help me, or know other way to use the algoritm for many g?

thanks