Question

it’s a simple issue about optim() function on R. I am trying to find the mimimum of a function depending on a loop’s variable.

f<-function(x,i){
  i*x
}

M<-matrix(0,10,1)

  for (i in 1:5) {
    M[i,1]<-optim(c(1),f,lower=-2,upper=2,method="L-BFGS-B")
  }
M

R answers me :

Erreur dans i * x : 'i' est manquant

I really don’t understand how to deal with this one….

Answer 1

You have to pass i to the optim call:

optim(1,f,i=i,lower=-2,upper=2,method="L-BFGS-B")

…and optim returns a list, so you need to pick which element in the list you want to store in M. I assume you want par:

M[i,1] <- optim(1,f,i=i,lower=-2,upper=2,method="L-BFGS-B")$par

UPDATE:

As Justin pointed out in the comments, the more R-like way to do this would be:

L <- list()
for (i in 1:5) {
    L[[i]] <- optim(1,f,i=i,lower=-2,upper=2,method="L-BFGS-B")
}
V <- sapply(L, `[[`, "par")