[Rd] quadprog

[Stefano Iacus]

Hy,
I have to solve the following minimization problem on x

Q(x) = 1/2 x' A x + x'b

where Q is a quadratic form and x is in the simplex (x_i>=0, sum_i x_i = 
1, i=1,...,n), A pos. def. and b a negative vector.

I have tried with quadprog routines but it gives me solutions of the form
x* = (...,1,..)
where the dots "." are zeroes "0".

the toolbox optim/quadprog in Matlab lead to the same results as 
R+quaprog package (they use the "same" qld algorithm).

Are there any more efficient methods based on lagrange multipliers + 
standard simplex minimization ? I think that the qld algorithm is rather 
general and then not efficient for my specific case.

Any idea or reference ?

Stefano








-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._