[R-sig-finance] Markowitz with non-linear constraints

[Molins, Jordi]

Dear all,

I am starting to use portfolio optimization with R. It seems pretty easy to
use Markowitz optimization with linear constraints, but I still have not
come up with a way to introduce non-linear constraints (e.g., \sum_i |w_i| =
1, to allow short positions to "burn" capital, or \sum_i w_i*w_i = constant,
in order not to have too much concentration in the portfolio). Is there an
easy way to introduce these non-linear constraints without resorting to
introduce lagrange multipliers, and then make these multipliers tend to
infinity? 

An additional question: is there a package that applies random matrix theory
to a correlation matrix? I would like to clean the correlation matrix from
spurious (low) eigenvalues, and not just deleting all small eigenvalues
without a clear-cut cut-off point.

Regards and thank you

Jordi



--------------------------------------------------------------------------------
The information contained herein is confidential and is inte...{{dropped}}