Optimization of portfolio Constituted from mutual funds of Tehran stock exchange using genetic algorithm

Expansion and growing complexity of financial markets, makes difficult Choosing assets for investors; Meanwhile, based on portfolio theory, diversification, while maintaining average returns, leads to lower fluctuations. Also, due to the complexity and speed of effective factors, constitution of optimal portfolio with Using traditional methods is difficult. This study is aimed at optimizing the portfolio consists of shares of mutual funds using genetic algorithm and comparing it with the traditional approaches and the impact of portfolio sizehas also been studied. For this purpose, the data of 30 mutual funds in the Iran's stock market from 2011 till 2013 are collected. Results indicate that genetic algorithms can be used to select a portfolio of shares of mutual funds. Using paired t-test determined that using genetic algorithms in portfolio selection is better than traditional methods. Also, the size of the portfolio did not influence the results and at all level, the genetic algorithm has better performance. Meanwhile, the larger and more diversified portfolio, genetic algorithm performance advantage over linear methods is more significant.