A hybrid framework integrating mathematical genetic Algorithm and Multi-Criteria decision making for robust Rule-Based stock market strategy optimization

Financial markets exhibit nonlinear dynamics, regime shifts, and conflicting performance objectives, making traditional optimization methods insufficient for robust trading strategy design. This study proposes a hybrid framework integrating a Mathematical Genetic Algorithm (MGA) with a Multi-Criteria Decision-Making (MCDM) method for evolving interpretable, rule-based trading strategies. The MGA employs mathematically governed operators and integer-constrained parameter evolution, while the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method evaluates candidates across return, risk, drawdown, and volatility. Using NIFTY NEXT 50 data and walk-forward validation, the proposed framework achieves superior risk-adjusted performance compared to a buy-and-hold benchmark, yielding higher annualized returns (0.165), higher Sharpe ratio (1.32), and lower maximum drawdown (0.080). Sensitivity analysis confirms robustness to variations in the MGA penalty parameter λ and TOPSIS weights. Findings demonstrate that embedding MCDM within MGA improves interpretability, multi-objective balance, and out-of-sample stability, offering a practical pathway for intelligent trading system design.