CODEWITHSUNDEEP
rloopsfor-loopmathematical-optimization
forecaster
forecaster

Reputation: 1159

looping in R efficent programming

I have the following program. I just want to run the for loop and save the results from each iteration into valg[i,], valo[i,] and obj[i,] matrix/table

I'm doing something incorrectly, I get the following error 

Error in `[<-`(`*tmp*`, i, , value = c(80.6413039191413, 40.0368791515846,  : 
  subscript out of bounds

This program requires DEoptim package

I would greatly appreciate any help 

g = matrix(c(1,6,1,1,1/6,
             1/6,1,6,1,1,
             1,1/6,1,6,1,
             1,1,1/6,1,6,
             6,1,1,1/6,1
             ), nrow=5, ncol=5)

a <- t(g)

lsm <- function(x){
  b <- a - outer(x,x,`/`)
  z <- sum(b^2)
  return(z)}

for (i in (1 : 10)){

outDEoptim <- DEoptim(lsm, lower = rep(0.001,nrow(a)), 
                      upper = rep(100,nrow(a)),DEoptim.control(trace = FALSE))

  valg[i,] = rbind(outDEoptim$member$bestmemit[200,]) ## Save for each itr i and input to optim below for each itr 

  lsmoptl <- optim(par=c(valg[i,]), fn=lsm, gr = NULL,
                 method = c("L-BFGS-B"))
  valo[i,] = lsmoptl$par/sum(lsmoptl$par) ## Save for each itr i 

  obj[i,] = lsmoptl$value ## Save for each itr i 
}

Upvotes: 0

Views: 170

Answers (1)

PascalVKooten
PascalVKooten

Reputation: 21433

Unless you forgot to post them and it is not the problem, you have to first initialize the object valg (and the others) for it to allow storing in indices.

For instance: 

valg <- matrix(1, nrow=10, ncol=3)

will initialize the object.

From then on you can store in it using a loop.

for (i in 1:10) {
    valg[i, ] <- 1:3
}

You may want to read some R tutorials online before posting elementary R issues.

Upvotes: 1

Related Questions