Investors’ behavior-based nonlinear goal programming models for financial portfolio selection considering future stock market trends

This is the authors' abstract. We don't add key points for this paper.

Financial markets are ever-changing, and portfolio selection (PS) problems frequently involve multiple optimization objectives in an imprecise investment environment. Real-world investors attempt to select a portfolio based on their preferences for each objective and their decision-making behaviors. Moreover, integrating investor behavior according to predicted future stock market trends results in more appropriate portfolios. To address such situations, we propose a multi-objective approach for modeling investors’ preferences regarding several objectives (portfolio return, risk, and beta) as well as their decision behaviors in relation to future market scenarios. Additionally, real-world features such as cardinality and diversification constraints were considered. This study proposes a novel behavior-based nonlinear goal programming model for PS that aligns with stock market trends. The concept of behavioral functions is introduced to model investors’ decisions in the portfolio optimization process. To clarify the effects of different risk attitudes on PS, we examined three commonly used types of risk: risk-averse, risk-neutral, and risk-seeking. Our approach is validated using the S&P 100, where three distinct investment scenarios (sideways, upward, and downward moving market trends) are analyzed separately. The proposed PS models are then applied to determine the best portfolios, demonstrating their effectiveness in formulating and integrating investor preferences and behaviors. • Propose novel behavior-based nonlinear goal programming (BBNGP) models for portfolio optimization. • Model investor preferences and behaviors informed by predicted market trends. • Develop a multi-objective portfolio selection framework under uncertainty and imprecise environments. • Address trade-offs between financial optimality and behavioral realism. • Validate the proposed models using S&P 100 data, including scenario analysis and comparative evaluation.