Multi-Asset Portfolio Optimization based on Conditional Value at Risk using Artificial Bee Colony Algorithm

Multi-asset portfolio management and optimization have always been of interest to investors. Due to the inflation in Iran market, different performance of the asset classes in different market conditions and the ability to earn more profit along with less risk by diversifying the types of assets, it seems necessary to select a portfolio consisting of stocks, foreign currency and commodities. In this paper, assets of the above categories, including Emami coins, American dollar, and 11 sector indices, are considered in the portfolio composition. Due to the importance of the risk measure in multi-asset portfolio optimization, a model with conditional value at risk, the historical simulation approach has been extended and its efficiency has been compared with the mean-variance model. The models have been solved using the artificial bee colony and imperialist competitive algorithms. The daily asset prices in the period 2013 to 2020 have been used to evaluate the models in Iran market. Results show that the mean-conditional value at risk model performs better than the mean-variance in the training and testing periods. Furthermore, optimized portfolios with the artificial bee colony algorithm could outperform the imperialist competitive algorithm based on the Sharpe ratio, conditional Sharpe ratio, and return on risk.