Investment portfolio selection using genetic algorithms

• Main Author: Zhang, Yawen
• Other Authors: Mao Kezhi
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Computer and Information Science; Engineering; Portfolio selection; Genetic algorithms; Multifactor quantitative model
• Online Access: https://hdl.handle.net/10356/176229
• Institution: Nanyang Technological University

The challenge of portfolio selection remains a critical focus within the domains of genetic algorithms and computational finance optimization. In the volatile financial markets, the objective of optimizing genetic algorithms (GAs) is to enhance the efficiency of portfolio decision-making processes and identify portfolios with superior risk-adjustment capabilities. This dissertation explores the application of an enhanced genetic algorithm, grounded in a multifactor quantitative model, for strategic portfolio decision-making to sift through multifaceted data and pinpoint stock portfolios that yield high returns and exhibit robust risk resistance. Initially,this dissertation constructs an individual fitness function calculation model incorporating multiple technical indicators based on the concept of multifactor quantification. Then, introduces industry diversification into the optimization process to prevent the algorithm from overly concentrating on local optima, thereby increasing the likelihood and efficiency of identifying global optima. Furthermore, it refines the genetic algorithm to overcome traditional methods’ drawbacks of high computational costs and premature convergence. Rigorous back-testing validates the effectiveness of this method, with comparative experiments between GA-utilized and non-GA-utilized models underscoring the superiority of the proposed approach, particularly during market downturns. The results affirm that GA-optimized portfolios achieve higher risk-adjusted returns and minimize drawdowns amidst market volatility. These insights highlight the potential of the optimized GA as a powerful tool for navigating complex market dynamics, offering investors a strategic advantage.