Back to feed
arXiv cs.LG
arXiv cs.LG
7/9/2026
Deep Reinforcement Learning for Reliability Based Bi-Objective Portfolio Optimization

Deep Reinforcement Learning for Reliability Based Bi-Objective Portfolio Optimization

Short summary

Researchers propose MORP-DRL, a deep reinforcement learning framework for multi-objective reliability-based portfolio optimization that jointly optimizes expected return and downside risk using variance, CVaR, and EVaR. The approach models heavy-tailed returns via GARCH(1,1), Extreme Value Theory, and t-copula structures, using a PPO-based strategy benchmarked against NSGA-II across pre-COVID, COVID, and post-COVID market regimes. Results show competitive risk-return performance with reduced downside risk during market stress and scalability to high-dimensional portfolios.

  • •MORP-DRL uses PPO to jointly optimize return and downside risk with variance, CVaR, and EVaR measures
  • •Asset returns modeled via GARCH(1,1), Extreme Value Theory, and t-copula with quasi-Monte Carlo scenario generation
  • •Benchmarked against NSGA-II across ten global equity indices in three market regimes, showing reduced downside risk during stress periods

Generated with AI, which can make mistakes.

#research-breakthrough#ai-tools
Read full article at arXiv cs.LG

Is this a good recommendation for you?

Comments

Failed to load comments. Please try again.

Explore more

Reinforcement Learning-Guided NSGA-II Enhanced with Gray Relational Coefficient for Multi-Objective Optimization: Application to NASDAQ Portfolio Optimization
arXiv cs.LG
The original title is "Some Large Language Models Exhibit Consistent Risk Attitudes"
arXiv CS.AI
Uncertainty-Aware Sequential Decision Rules for Event-Triggered LLM Invocation in Streaming Systems
arXiv cs.LG
Out-of-Distribution Generalization of Risk Aversion in Language Models
arXiv cs.LG