Optimizing Portfolios with Modified NSGA-II Solutions

In financial management, portfolio optimization remains a central challenge. From Markowitz’s Mean-Variance Optimization to modern deep learning methods, capturing the multi-faceted nature of portfolio decisions is complex. This paper emphasizes the potential of Multi-objective Evolutionary Algorithms (MOEAs), specifically the NSGA-II. We propose key modifications to NSGA-II, including refined selection strategies, dynamic mutation parameters and optimization for initialization. Initial results showcase enhanced performance and a more comprehensive Pareto front. Additionally, we employ the Monte Carlo Markov Chain (MCMC) for a ten-year portfolio projection. Future avenues include comparisons with other algorithms and expanding our refined NSGA-II’s application to diverse optimization tasks.