A New Efficient Metaheuristic Model for Stock Portfolio Management and its Performance Evaluation by Risk-adjusted Methods

In this research, we proposed a new metaheuristic technique for stock portfolio multi-objective optimization employing the combination of Strength Pareto Evolutionary Algorithm (SPEA), Adaptive Neuro-Fuzzy Inference System (ANFIS) and Arbitrage Pricing Theory (APT). To generate the more precise model, ANFIS has implemented to envisage long-term movement values of the Tehran Stock Exchange (TSE) indices including TSE TEPIX and TSE TEDIX making use of technical indicators. The SPEA is exerted to choose several characteristics of technical indicators that these types of chosen essential characteristics strengthen the overall performance of the forecasting model. This research applied the suggested model in Tehran Stock Exchange. The research sample contains panel data for top 50 Companies of Tehran Stock Exchange over a ten-year interval from 2007 to 2017. The efficient procedures on actual market information are examined and explain the performance of the offered model under true limitations from the experiential assessments; we clearly discover that SPEA-ANFIS-APT forecasting technique considerably performs better than the other portfolio optimization models. The suggested hybrid optimization approach provides considerable enhancements and also innovation in the portfolio management and investment strategies under unpredictable and uncertain stock exchange without human interference, with a diversification procedure, thereby supplying satisfactory and ideal returns with minimum risk. Furthermore, the planned portfolio model SPEA-ANFIS-APT attains appropriate and acceptable functionality among diverse portfolio models despite oscillations in a stock exchange conditions. In comparison with the outcomes of various other approaches, the supremacy of the offered model is approved.