Developing Meta-heuristic AntLion-Genetic and PBILDE Algorithms to Portfolio Optimization in Tehran Stock Exchange

In financial studies, portfolio can be defined as a set of investments that are selected and accepted by an individual or institution. Portfolio selection is one of the main concerns of investors in financial markets. The average-variance model with bound restrictions is considered as one of the main models in solving the portfolio optimization problem. In terms of complexity, this model is a polynomials NP-hard non-linear problem that cannot be accurately solved. In this study, an Antlion optimizer- Genetic algorithm (ALOGA) and a population based incremental learning and differential evolution algorithm (PBILDE), which are modern meta-heuristic models for solving optimization problem, are used to optimize the investment portfolio through increase the return and reduce the risk. Among 591 companies listed on Tehran stock exchange from April 2012 through March 2015, 150 companies were selected as the final sample using screening method. The data of these companies were analyzed using the applied algorithms in this research and their efficiency was compared together. The results indicate that ALOGA and PBILDE algorithms both are suitable for solving the portfolio optimization problem. In addition, using the ALOGA algorithm, it is possible to create an optimal portfolio with high accuracy and efficiency.