Diffusion-Augmented Reinforcement Learning for Robust Portfolio Optimization under Stress Scenarios

Himanshu Choudhary, Arishi Orra, Manoj Thakur

Keywords: Portfolio Optimization, Deep Reinforcement Learning, Generative Models, Quantitative Finance, Scenario Generation

TL;DR: The approach consists of diffusion-based data augmentation for crisis scenarios, a custom RL environment, and an iterative training procedure.

Abstract: In the ever-changing and intricate landscape of financial markets, portfolio optimisation remains a formidable challenge for investors and asset managers. Conventional methods often struggle to capture the complex dynamics of market behaviour and align with diverse investor preferences. To address this, we propose an innovative framework, termed Diffusion-Augmented Reinforcement Learning (DARL), which synergistically integrates Denoising Diffusion Probabilistic Models (DDPMs) with Deep Reinforcement Learning (DRL) for portfolio management. By leveraging DDPMs to generate synthetic market crash scenarios conditioned on varying stress intensities, our approach significantly enhances the robustness of training data. Empirical evaluations demonstrate that DARL outperforms traditional baselines, delivering superior risk-adjusted returns and resilience against unforeseen crises, such as the 2025 Tariff Crisis. This work offers a robust and practical methodology to bolster stress resilience in DRL-driven financial applications.

Submission Number: 76