Mon, Nov 17, 2025
[Archive]
Volume 17, Issue 3 (7-2025)
itrc 2025, 17(3): 58-69 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Sharbaf Movassaghpour S, Kargar M, Bayani A, Assadzadeh A, Khakzadi A. Adaptive Portfolio Optimization with Multi-Agent Deep Reinforcement Learning and Short-Term Performance Analysis. itrc 2025; 17 (3) :58-69
URL: http://ijict.itrc.ac.ir/article-1-738-en.html
URL: http://ijict.itrc.ac.ir/article-1-738-en.html
1- Department of Computer Engineering, Islamic Azad University, Tabriz Branch Tabriz, Iran
2- Department of Computer Engineering, Tabriz Branch, Islamic Azad University, Tabriz, Iran ,kargar@iaut.ac.ir
3- Department of Computer Engineering, Tabriz Branch, Islamic Azad University, Tabriz, Iran
4- Department of Computer Engineering Islamic Azad University, Tabriz Branch Tabriz, Iran ali.khakzadi@iau.ir
2- Department of Computer Engineering, Tabriz Branch, Islamic Azad University, Tabriz, Iran ,
3- Department of Computer Engineering, Tabriz Branch, Islamic Azad University, Tabriz, Iran
4- Department of Computer Engineering Islamic Azad University, Tabriz Branch Tabriz, Iran ali.khakzadi@iau.ir
Abstract: (643 Views)
This research presents a novel portfolio optimization framework using deep reinforcement learning (DRL).
Traditional methods rely on static models or single-agent strategies, which struggle with market dynamics. We propose
a dynamic system to address this by selecting the best-performing DRL agent based on recent market conditions. The
framework evaluates five DRL agents, A2C, SAC, TD3, DDPG, and PPO, allocating portfolio weights based on short-
term performance. A selection mechanism identifies the top agent using cumulative returns over the prior ten days,
leveraging multiple agents' strengths. This adaptive approach embraces the philosophy that no single strategy
consistently outperforms in all market conditions, making flexibility and continuous learning essential for robust
portfolio management. Backtesting on Dow Jones data shows our method enhances cumulative returns and risk-
adjusted performance, achieving an 11.43% average annual return, 38.29% cumulative returns, and a 0.832 Sharpe
ratio, outperforming individual DRL agents.
Traditional methods rely on static models or single-agent strategies, which struggle with market dynamics. We propose
a dynamic system to address this by selecting the best-performing DRL agent based on recent market conditions. The
framework evaluates five DRL agents, A2C, SAC, TD3, DDPG, and PPO, allocating portfolio weights based on short-
term performance. A selection mechanism identifies the top agent using cumulative returns over the prior ten days,
leveraging multiple agents' strengths. This adaptive approach embraces the philosophy that no single strategy
consistently outperforms in all market conditions, making flexibility and continuous learning essential for robust
portfolio management. Backtesting on Dow Jones data shows our method enhances cumulative returns and risk-
adjusted performance, achieving an 11.43% average annual return, 38.29% cumulative returns, and a 0.832 Sharpe
ratio, outperforming individual DRL agents.
Keywords: Deep Reinforcement Learning, Consulting System, Portfolio Optimization, Tactical Decision, Cumulative Return
Type of Study: Research |
Subject:
Information Technology
Send email to the article author
| Rights and permissions | |
 | This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
