A0828
Title: A many-objective evolutionary algorithm for a portfolio optimization problem with ESG and diversification goals
Authors: Massimiliano Kaucic - University of Trieste (Italy) [presenting]
Abstract: A novel portfolio design is introduced that extends the Markowitz framework by considering objectives to optimize the portfolio diversification and the sustainability value of the investment in addition to portfolio mean and variance. Some real-world constraints are further included, namely full-investing, buy-in thresholds, and portfolio size. A linear combination of excess kurtosis and squared skewness is used for the diversification function. In contrast, the sustainable objective function is defined as the weighted sum of the relative sustainability values of the portfolio constituents. The resulting four-objective mixed-integer portfolio optimization problem has the following computational challenges: the presence of nonlinear objective functions and discrete constraints. A reference-point-based many-objective evolutionary algorithm is developed following the NSGA-III framework to tackle this optimization problem and perform a set of numerical experiments to test its effectiveness. Finally, the ex-post features of the portfolios are analyzed on the Pareto front.