A Hybrid Multi-Objective Optimization Approach For Portfolio Selection Problem

Portfolio selection problem is a major subject in finance where investors deal with selecting satisfying portfolio which is composed of a vast number of risky assets, under some restricting criteria that are defined by themselves. Asset prices can be effected from different events, such as political crisis, financial turmoil and technological improvements. Due to uncertainty nature of these events, it is difficult to forecast future prices of assets. However, Markowitz's Modern Portfolio Theory, which is mainly focused on portfolio risk, introduced a new idea for asset diversification in portfolio optimization. According to this approach, an investor can reduce portfolio risk simply by holding combinations of assets that are not perfectly positively correlated and also efficient portfolio can only be obtained by focusing portfolio return and risk together. In this paper, a two stage multi objective portfolio selection model is proposed for obtaining best portfolio. In the first stage, Pareto efficient portfolios are obtained by genetic algorithm with using mean and variance of assets. Then in the second stage a multi criteria decision method is applied for ranking Pareto-optimum portfolios that are obtained in previous stage. Effectiveness of criteria, such as entropy measures and higher moments are taken into consideration and also performance ratios are examined in evaluating Pareto efficient portfolios and their rankings. An illustrated example is given and results of proposed model are discussed in experimental section