Publications

Journal Articles

HMU: Heuristic Memory Utilization for Multi-Agent Reinforcement Learning

Published in IEEE Transactions on Cognitive and Developmental Systems, 2026

In cooperative Multi-Agent Reinforcement Learning (MARL), agents attempt to complete the same task, such as defeating enemies and scoring goals. Due to the trial-and-error mechanism, MARL algorithms usually require a large amount of interactions with the environment to achieve convergence. .

Recommended citation: R. Tang et al., "HMU: Heuristic Memory Utilization for Multi-Agent Reinforcement Learning," in IEEE Transactions on Cognitive and Developmental Systems, doi: 10.1109/TCDS.2026.3685230.
Download Paper | Download Bibtex

Conference Papers

Autonomous Partner Selection for Cooperative Multi-Agent Reinforcement Learning

Published in AAAI-26 Poster, 2026

In cooperative Multi-Agent Reinforcement Learning (MARL), the subgroup-wise learning is employed to assign sub-tasks to agents towards the enhancement of team collaboration.

Recommended citation: Rui Tang, Biao Luo, Yongzheng Cui. " Autonomous Partner Selection for Cooperative Multi-Agent Reinforcement Learning" (2026).

Beyond Monotonicity: Revisiting Factorization Principles in Multi-Agent Q-Learning

Published in AAAI-26 oral, 2026

Value decomposition is a central approach in multi-agent reinforcement learning (MARL), enabling centralized training with decentralized execution by factorizing the global value function into local values.

Recommended citation: Tianmeng Hu, Yongzheng Cui, Rui Tang, Biao Luo, Ke Li. (2026). "Beyond Monotonicity: Revisiting Factorization Principles in Multi-Agent Q-Learning." AAAI2026.
Download Paper