The field of multi-agent reinforcement learning (MARL) is rapidly advancing, with a focus on developing more efficient, scalable, and robust methods for coordinating agent behavior. Recent work has emphasized the importance of communication, cooperation, and adaptability in MARL systems, particularly in complex, dynamic environments. Researchers are exploring new approaches to address challenges such as partial observability, limited communication, and conflicting objectives. Notable developments include the use of hierarchical frameworks, graph-based methods, and decentralized control strategies. These innovations have the potential to improve performance in a wide range of applications, from traffic control and autonomous vehicles to smart grids and healthcare systems. Noteworthy papers include: Scalable Population Training for Zero-Shot Coordination, which proposes an efficient training framework for zero-shot coordination in MARL. HCPO: Hierarchical Conductor-Based Policy Optimization in Multi-Agent Reinforcement Learning, which introduces a conductor-based joint policy framework for cooperative MARL. Transformer-Based Scalable Multi-Agent Reinforcement Learning for Networked Systems with Long-Range Interactions, which presents a unified transformer-based MARL framework for modeling long-range dependencies in networked systems.
Advances in Multi-Agent Reinforcement Learning
Sources
Robust and Efficient Communication in Multi-Agent Reinforcement Learning
Scalable Population Training for Zero-Shot Coordination
Deviation Dynamics in Cardinal Hedonic Games
Aspiration-based Perturbed Learning Automata in Games with Noisy Utility Measurements. Part A: Stochastic Stability in Non-zero-Sum Games
Convergence of Multiagent Learning Systems for Traffic control
Goal-Oriented Multi-Agent Reinforcement Learning for Decentralized Agent Teams
HCPO: Hierarchical Conductor-Based Policy Optimization in Multi-Agent Reinforcement Learning
Resilient and Efficient Allocation for Large-Scale Autonomous Fleets via Decentralized Coordination
Conditional Diffusion Model for Multi-Agent Dynamic Task Decomposition
Transformer-Based Scalable Multi-Agent Reinforcement Learning for Networked Systems with Long-Range Interactions
Quantifying Distribution Shift in Traffic Signal Control with Histogram-Based GEH Distance
Fair-GNE : Generalized Nash Equilibrium-Seeking Fairness in Multiagent Healthcare Automation
Z-Merge: Multi-Agent Reinforcement Learning for On-Ramp Merging with Zone-Specific V2X Traffic Information
Data-driven control of network systems: Accounting for communication adaptivity and security
Symmetry-Breaking in Multi-Agent Navigation: Winding Number-Aware MPC with a Learned Topological Strategy
Decentralized Gaussian Process Classification and an Application in Subsea Robotics
A Scenario Approach to the Robustness of Nonconvex-Nonconcave Minimax Problems