Stock portfolio optimization using hill climbing and simple human learning optimization algorithms as a decision...

Area of Science:

• Quantitative Finance
• Computational Finance
• Investment Management

Background:

• Stock portfolio optimization aims to maximize returns while minimizing risk.
• Traditional methods may not fully capture complex market dynamics.
• Integrating fundamental analysis offers a robust approach to stock selection.

Purpose of the Study:

• To develop a decision support system (DSS) for stock portfolio optimization.
• To hybridize nature-inspired algorithms (hill climbing and SHLO) with fundamental analysis modules.
• To evaluate the performance of optimized portfolios against various risk profiles and market indices.

Main Methods:

• Developed a DSS integrating intrinsic value and financial health analysis modules.
• Designed custom datasets using historical fundamental stock data.
• Employed a novel fitness function with hill climbing and SHLO algorithms for optimization.

Main Results:

• Optimized portfolios showed decreasing returns from 55% to 24% as risk tolerance increased.
• Increasing portfolio cardinality led to decreased returns.
• Fundamentally undervalued portfolios demonstrated superior performance compared to growth portfolios.

Conclusions:

• The developed DSS effectively optimizes stock portfolios based on fundamental analysis.
• Optimized portfolios consistently outperformed US market indices (>80% of the time).
• Hybridizing nature-inspired algorithms enhances portfolio selection accuracy and risk management.