Hello,
I am trying to optimise a nonlinear model to derive 'best-fit' parameter
esimates using the genoud function. I have been using the genetic algorithm
- gafit - in order to do this, but I am getting parameter estimates that do
not always reach the global minimum. I am very keen to apply genoud to
optimising this model to see if my results will improve, and also out of
personal interest. However the problem I am having is, if I have a set of
'best-fit' parameters (a,b,c) and corresponding variables (X1,X2,X3) for my
function, how does one supply the variables to the function that is
optimised in genoud?
i.e.
WSSR <- function(B,D) {
a<-B[1]
b<-B[2]
c<-B[3]
Y <- D[1]
X1 <- D[2]
X2 <- D[3]
X3 <- D[4]
chi2 = (Y - a*X1+b*X2-c*X3)^2
return(chi2)
}
genoud(WSSR, nvars=3, pop.size=5000, max=FALSE)
Genoud optimises the first variable (in this case the vector B) supplied to
the function, however how do I pass the data (the matrix D) to the WSSR
function within the genoud framework? In the optim function, you can
explicitly define any variables within the function framework, i.e.
optim(WSSR, D=data, tol=0.001, .... )
Passing it this way does not work in genoud, and I am at loss as to how I
would pass data to the function to be optimised. To put it simply, how do
you perform a nonlinear least squares optimisation using genoud?
Best Regards
--
======================================
Rhys Whitley
PhD Candidate
Institute for Water and Environmental Resource Management
Department of Physics and Advanced Materials
University of Technology, Sydney
Australia
======================================
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
