Optimizing portfolio of SET stocks using convex quadratic programming

This article considers the portfolio selection problem of 37 leading stocks in the Stock Exchange of Thailand. We use the “expected returns-variance of returns” rule proposed by Markowitz, which formulates the problem as a convex quadratic program. To estimate the expected returns and variance of returns, we propose to use linear model of historical daily average price of each security. Finally, we implement the algorithm in Matlab and show the results it computes for different parameters.