Reinforcement learning-driven optimal capital allocation for intelligent finance

Optimal capital allocation in financial markets has long been a challenging task due to the dynamic nature of market conditions, risk factors, and the complexity of financial data. Traditional methods of portfolio management often fail to adapt to market shifts and do not fully leverage potential of modern machine learning techniques. To develop a model that dynamically optimizes capital allocation, leveraging reinforcement learning (RL), with focus on risk management, stability, and adaptability. Proposed model, Dynamic Actor-Critic (DAC)-enriched Clipped Proximal Policy Optimization (DAC-CPPO) model for portfolio optimization. A Dynamic Actor-Critic (DAC)-enriched Clipped Proximal Policy Optimization (DAC-CPPO) model is developed for portfolio optimization. The Dynamic Actor-Critic module learns optimal portfolio policies by evaluating actions and maximizing long-term returns, while the CPPO mechanism ensures stable policy updates and risk-controlled portfolio optimization. The Intelligent Finance Assets dataset contains 7,000 records with 32 features, whereas the Massive Yahoo Finance dataset includes 603k records with 9 features, collected from publicly available Kaggle sources for financial analysis. Normalization and data cleaning techniques for smooth data integration and model readiness, and LDA extracts discriminative financial features to improve portfolio allocation prediction performance. Experimental settings are implemented using Python. Proposed DAC-CPPO model, achieving 97.6% accuracy, 0.074 MAE, 0.081 RMSE, 1.91 Sharpe ratio, 1.12 Cumulative return, and 0.14 Volatility, demonstrates superior risk-adjusted performance, offering higher returns while managing market risks effectively compared to other existing models. The DAC-CPPO model offers a robust solution for practical capital allocation by combining advanced RL techniques, ensuring effective decision-making while managing risks in dynamic financial environments.