Investment portfolio optimization with transaction costs through a multiobjective genetic algorithm: an applied case to the Colombian Stock Exchange

This paper discusses portfolio optimization by considering constraints imposed by financial markets and conditions of projects with excess liquidity, such as transaction costs, limited budget and short time planning horizons. In light of these constraints, conventional models are found to generate non-efficient portfolios. Consequently, a mathematical model is formulated and a multiobjective genetic algorithm is implemented in order to find efficient portfolios in the Colombian Stock Exchange (Bolsa de Valores de Colombia), minimizing risks and maximizing profits. In addition, results are shown which allow comparison between those portfolios obtained through the proposed model and the mean-variance model, highlighting the importance of transaction costs and budget in investment decision making