Multistage, Multiswarm Particle Swarm Optimization for Investment Portfolio Selection

Abstract

This study proposes a new metaheuristic-based approach, identified in the literature as MS2PSO, to solve the problem of investment portfolio selection in the Brazilian stock market. The study was divided into two experiments that evaluated the quality of the solutions obtained by the algorithm and the effectiveness of the recommended portfolios in different time windows and risk profiles. The results indicated that MS2PSO presented slower convergence but offered more satisfactory results compared to PSO. In addition, the portfolios recommended by the proposed method showed positive and negative performances according to the risk profile, and all of them outperformed the benchmark in terms of the overall highest gain obtained during the period. This study contributes to the development of the area of finance and economics by providing a sophisticated and efficient solution for investment portfolio selection.