CODEWITHSUNDEEP
rfunctionoptimizationvectormaximize
ES1099
ES1099

Reputation: 1

Optimizing a function in two variables, which creates a vector as an output?

Background: We run a simulation which creates a matrix as an output. This output is then put into a function (with two variables) which creates a matrix/vector as an output.

Is there an easy way to optimize the function's parameters so that the output vector contains the maximum output possible?

This produces the simulated matrix:

sim=1000 
subp=12
set.seed(1)
eps.m <- matrix(rnorm(sim*subp), sim, subp, byrow=TRUE)

mü5=0.0003934473
sigma5=0.07147787

r.m<-mü5-sigma5^2/2 + sigma5*eps.m 

S.m <- matrix(0,sim,subp+1)
S.m[,1] <- 1000 

for (j in 2:13) S.m[,j]<-S.m[,j-1]*exp(r.m[,j-1])

Some variables like "mü5", "sigma5" are defined a bit before the simulation, if you need them please let me know.

And this is the mentioned function dependent on "mu" and "b":

CPPIopt <- function(mu,b) {
  
  FT <- 650000
  W0 <- 1000000
  T0 <- 1
  m <- 12
  rf <- log(1.01)
  
  SIM<-matrix(0,nrow(S.m),m+1)
  
  SIM[,1]       <- W0
  Tt.T[1]       <- T0
  for(i in 2:(m+1)) {
    
    Wt          <- SIM[,i-1]
    Tt.T[i]     <- Tt.T[i-1]-(T0/m)
    Ft[i]       <- FT*exp((-rf)*Tt.T[i])
    Ct[i]       <- max(Wt-Ft[i], 0)
    Xr.t[i]     <- min(mu*Ct[i],b*Wt)
    Xf.t[i]     <- W0-Xr.t[i]
    SIM[,i]     <- (Xr.t[i]*S.m[,i])/S.m[,i-1] + Xf.t[i]*exp(rf/m)}
    
  return(SIM[,13])
  }

Unfortunately I'm a beginner in R and it's my first post, so please have mercy with me.

Any kind of help or information to the problem would be greatly appreciated, big thanks in advance!

Upvotes: 0

Views: 100

Answers (0)

Related Questions