Portfolio optimization with linear programming approach

Abstract

Variance has been commonly used as risk measure in portfolio optimization since the introduction of mean-variance model by Markowitz. However, the mean-variance model is a quadratic programming model. It takes longer time to solve this model. Some linear programming models have been proposed to overcome the disadvantage of mean-variance model. The purpose of this paper is to compare the portfolio composition and performance of different linear programming models with mean-variance model. The linear programming models that will be discussed in this paper are mean absolute deviation, minimax and conditional value at risk.