Genetic algorithms for portfolio selection problems with non-linear objectives

摘要

Genetic algorithm is proven to lead to better solutions in solving combinatorial optimization problems like portfolio selection. Generally, investors in portfolio selection, simultaneously consider such contradictory objectives as the rate of return, risk and liquidity. We employed genetic algorithm (GA) model to select the best portfolio in 50 supreme Tehran Stock Exchange companies in order to optimize the objectives of the rate of return, systematic and non-systematic risks, return skewness, liquidity and Sharpe ratio. Finally, the obtained results were compared with the results of Markowitz's classic model. The comparison indicated that although, the rate of return of the portfolio of GA model was less than that in the Markowitz’s classic model, GA had basically some advantages in decreasing risk in the sense that it completely covers the rate of return and leads to better results and proposes more versatility portfolios when compared with the other models. Therefore, it could be concluded that as far as selection of the best portfolio is concerned, GA model can lead to better results and may help the investors to make the best portfolio selection.