A Comparative Study of Metaheuristics Techniques for Portfolio Selection Problem

Portfolio Selection Problem (PSP) is one of the major interesting research areas in finance which have drawn interest of several researchers over the years. Over time, the different approaches had been engaged in solving the PSP ranging from computational techniques to metaheuristics techniques with varying results. In this study, we engaged three different metaheuristics techniques under this same condition to solve extended Markowitz mean-variance portfolio selection model. The three metaheuristics techniques are Non-dominated Sorting Genetic Algorithm II (NSGAII), Speed-constrained Multi-objective Particle Swarm Optimization (SMPSO) and Generalized Differential Evolution 3 (GDE3). A comparative analysis was carried out with results obtained with existing benchmark data available in literature. The outcome of the findings reveals that SMPSO shows superior performance, followed by NSGAII in many different instances, however, the mean execution time of GDE3 was the fastest among the three techniques considered