Robust Multi -Objective Portfolio Optimization with ENS - NSGA -II Under Real -World Constraints

This study explores portfolio optimization by employing the Enhanced Non -dominated Sorting Genetic Algorithm II (ENS -NSGA -II), a robust multi -objective evolutionary algorithm, alongside classical optimization methods. Utilizing data from ۱۳۵ companies listed on the Tehran Stock Exchange over the period ۲۰۱۳ –۲۰۲۴, an extended Markowitz framework was constructed based on return and semi -variance as key performance criteria. To better reflect market realities, several practical investment constraints were incorporated, transforming the problem into a multi -objective optimization task. The ENS -NSGA -II algorithm, equipped with adaptive mutation control and improved diversity preservation mechanisms, significantly outperformed classical models by generating portfolios with superior return -risk trade -offs under nonlinear and volatile market conditions. Furthermore, the algorithm exhibited strong convergence behavior and stability across multiple independent runs and time intervals. These findings confirm that ENS -NSGA-II offers a powerful and flexible approach for constructing efficient investment portfolios in uncertain financial environments, providing investors with a reliable decision -making tool that adapts to dynamic market complexities.

کلیدواژه ها:

Risk ، Return ، Portfolio ، Classical Optimization ، Intelligent Optimization