Evolutionary 4-Objective Optimization Portfolio Algorithms for fuzzy and non-fuzzy selection
Subject Areas : Financial engineeringMohammad javad Salimi 1 , Mohammad Taghi Taqhavi Fard 2 , Mirfeiz Fallahshams 3 , Hadi Khajezadeh Dezfuli 4
1 - Assistant Prof., Faculty of Management and Accounting, Allameh Tabataba'i University, Tehran, Iran
2 - Associate Prof., Faculty of Department of Industrial Management, Allameh Tabataba'i University, Tehran, Iran
3 - Associate Prof., Financial management Department, Azad University, Tehran, Iran
4 - PHD candidate of Finance at Allameh Tabatabai University(ATU),Tehran, Iran.
Keywords: fuzzy logic, Modern Portfolio Theory, Post Modern Portfolio Theory, Financial Modeling, Optimize Portfolio Selection, Evolutionary Multi-Objective Algorithm,
Abstract :
In choosing the optimal portfolio, we must consider various criteria, some of which are determined by the nature of the optimization and some are determined by the investor's desire. Therefore, in this paper, multi-objective optimization models are designed and solved in MATLAB software environment. These models are designed in such a way that both the nature of the portfolio optimization, the considerations of the investor and the uncertain nature of the future return on assets, are taken into account. After designing the models in fuzzy and non-fuzzy (simple) conditions, due to their NP-HARD nature, a dedicated NSGA-II algorithm was used to solve it. After solving the models, the best portfolio from attained Pareto frontier, based on the Sortino ratio, be chosen. After that all of the obtained portfolios are compared according to the Treyner ratio. The results of statistical tests clearly show that the proposed models have a high power in choosing portfolios with maximum returns and a minimum risk. The results also indicate that that the designed models, with use of fuzzy logic in quadratic models creates more favorable results than simple models without using possibility theory and fuzzy logic.
_||_
