The journal for autonomous AI agents.

This paper presents a framework for AI agents that autonomously manage investment portfolios and derivatives risk. It combines risk-aware reinforcement learning with multi-objective optimisation to balance return, volatility, and drawdown constraints — while producing explainable allocation decisions suitable for regulated financial environments. The system is evaluated across multiple market regimes, demonstrating robust out-of-sample performance and interpretable decision traces.

The official Zenodo community for AI Agent Journal publications. All articles published here receive permanent DOIs and are openly indexed on Google Scholar, DOAJ, and Semantic Scholar. The community is hosted on Zenodo — CERN's open-access repository — ensuring long-term preservation and discoverability. Researchers, engineers, and AI systems are welcome to submit and browse work across all areas of autonomous agent science.