Skip to content

Rglpk_solve_LP

2 messages · Subhrangshu Nandi, Guy Yollin

Original
Guy Yollin
# 
Nandi,

Rglpk_solve_LP is simply a general-purpose linear program solver.

Some portfolio optimization problems like conditional value-at-risk (CVaR) optimization or mean absolute deviation (MAD) optimization can be formulated as linear programs and solved using an LP solver like Rglpk_solve_LP.

To understand how this is done I would highly recommend texts like:
  Introduction to Modern Portfolio Optimization with NuOPT, Bernd Scherer, Doug Martin, 2007
  Portfolio Construction and Risk Budgeting, Bernd Scherer, 2007

In these types of optimization problems, the entire returns matrix (which intrinsically includes all of the information regarding covariance) is typically passed to linear programming solver.

Best,

Guy


  


-----Original Message-----
From: r-sig-finance-bounces at stat.math.ethz.ch [mailto:r-sig-finance-bounces at stat.math.ethz.ch] On Behalf Of Subhrangshu Nandi
Sent: Tuesday, May 26, 2009 12:17 PM
To: r-sig-finance at stat.math.ethz.ch
Subject: [R-SIG-Finance] Rglpk_solve_LP

When you are trying to solve a portfolio optimization problem using the
package Rglpk_solve_LP, how do you put in the covariance matrix of the
assets? From some of the examples that are available online, it looks like
it can only be used in portfolios where the assets are uncorrelated. I may
be incorrectly interpreting the features of the package. Any thoughts?

Thanks,
-Nandi.