Portfolio Selection Problem Using CVaR Risk Measures Equipped with DEA, PSO, and ICA Algorithms

Investors always pay attention to the two factors of return and risk in portfolio optimization. There are different metrics for calculation factor, among which most important one is Conditional Value at Risk (CVaR). On other hand, Data Envelopment Analysis (DEA) can be used form optimal evaluate its efficiency. In these models, created by stocks or companies with high Since search space vast actual markets there limitations such as number assets their weight, optimization problem becomes difficult. Evolutionary algorithms a powerful tool deal difficulties. The automotive industry Iran involves international manufacturers. Hence, it essential investigate market related this invest it. Therefore, study we examined based on price index group, then optimized using methods. first method, CVaR measurement was modeled means DEA, Particle Swarm Optimization (PSO) Imperial Competitive Algorithm (ICA) were solve proposed model. second PSO ICA applied model, efficiency portfolios analyzed. Then, methods compared classic Mean-CVaR results showed that skewed right, possibility an increase return. Most favorable This displayed produced DEA-Mean-CVaR model increased because investment proposal basedon stock highest expected effective three levels. It found when solving evolutionary algorithms, decreased. efficient boundary algorithm higher than algorithm, more portfolios.Therefore, successful optimizing portfolio.