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

The goal of this research is to develop a decision support system for stock portfolio optimization using hill climbing and SHLO algorithms based on fundamental analysis of stocks. Portfolio optimization involves constructing a portfolio that maximizes returns while minimizing risk. The novelty in methodology is 'hybridizing' nature-inspired algorithms for optimized portfolio selection with two independent modules: intrinsic value of stocks and financial health analysis. This integrated approach aids decision-making by considering multiple dimensions of stock performance. Custom datasets are designed for each input module using historical fundamental data. The DSS output presents an optimized portfolio. Comparison for different risk profiles shows that as risk increases, returns of optimized portfolios decrease from 55 % to 24 %. Results for keeping other inputs the same for varying cardinality show that as cardinality increases, returns decrease. The results show that fundamentally undervalued portfolios outperform growth portfolios by a considerable margin. We conclude that optimized portfolios with varying constraints, >80 % of the time, outperform US market indices. Key contributions include:•Developed a decision support system using intrinsic value and financial health analysis.•Novel fitness function for optimization using hill climbing and SHLO.•Integrated module outputs with hill climbing and SHLO for portfolio optimization.