Stock Portfolio Optimization with Using a New Hybrid Evolutionary Algorithm Based on ICA and GA: Recursive-ICA-GA (Case Study of Tehran Stock Exchange)

Abstract

How to allocate resources and select the type of investment is very important . The optimal allocation, especially in the financial markets of countries that are paced growth factor, is very significance. In this research toward optimizing resource allocation, an innovative learning algorithm will used to select and optimize portfolio in Tehran Stock Exchange. a new method is proposed based on the combination of ICA (Imperial Competitive Algorithm) and GA (Genetic Algorithm) which improves the convergence speed and accuracy of the optimization results. The new algorithm, which is named R-ICGA (Recursive- ICA-GA), runs ICA and GA consecutively. It is shown that a fast decrease occurs while the proposed algorithm switches from ICA to GA. The main goal of the new proposed algorithm is to achieve a faster optimization technique by applying this fast decrease. Moreover, the simple combination of ICA and GA, which is named ICA-GA, is presented in this study. These two combination schemes of ICA and GA are used for comparing with other conventional algorithms. Finally, three fitness functions are used for comparing the suggested algorithms. The obtained results show that compared with the previous method, the proposed algorithms are at least 32% faster in optimization processes; also the variance convergence speed is smaller than the ICA and GA.