Mean Field Reinforcement Learning

Mean-field reinforcement learning (MFRL) tackles the scalability challenges of multi-agent reinforcement learning by approximating large populations of interacting agents with a single representative agent interacting with an average population state. Current research focuses on improving the accuracy and computational tractability of MFRL algorithms, including model-based approaches that incorporate uncertainty and decentralized methods that allow for heterogeneous agents and local observations. This framework offers significant potential for addressing large-scale problems in areas like traffic control, robotics, and resource management, where traditional multi-agent methods struggle due to computational complexity. The development of robust and efficient MFRL algorithms is a key area of ongoing investigation.

Papers

October 23, 2024

Scalable Offline Reinforcement Learning for Mean Field Games
Axel Brunnbauer, Julian Lemmel, Zahra Babaiee, Sophie Neubauer, Radu Grosu
Offline Reinforcement Learning Robust Policy Mean Field Game Mean Field Reinforcement Learning

May 1, 2024

MF-OML: Online Mean-Field Reinforcement Learning with Occupation Measures for Large Population Games
Anran Hu, Junzi Zhang
Multi Agent Reinforcement Learning Multi Agent Mean Field Game Approximate Nash Equilibrium Occupation Similarity Large Scale Game Mean Field Reinforcement Learning

February 8, 2024

When is Mean-Field Reinforcement Learning Tractable and Relevant?
Batuhan Yardim, Artur Goldman, Niao He
Multi Agent Reinforcement Learning Mean Field Relative Relevance Bounded Rational Mean Field Reinforcement Learning

June 29, 2023

Safe Model-Based Multi-Agent Mean-Field Reinforcement Learning
Matej Jusup, Barna Pásztor, Tadeusz Janik, Kenan Zhang, Francesco Corman, Andreas Krause, Ilija Bogunovic
Safe Multi Agent Reinforcement Learning Vehicle Trajectory Data Mean Field Reinforcement Learning

May 18, 2023

On the Statistical Efficiency of Mean-Field Reinforcement Learning with General Function Approximation
Jiawei Huang, Batuhan Yardim, Niao He
Optimal Policy Mean Field Sample Efficiency Mean Field Game \Epsilon$ Nash Equilibrium Mean Field Reinforcement Learning

April 25, 2023

Partially Observable Mean Field Multi-Agent Reinforcement Learning Based on Graph-Attention
Min Yang, Guanjun Liu, Ziyuan Zhou
Multi Agent Reinforcement Learning Graph Attention Multi Agent Reinforcement Mean Field Reinforcement Learning

May 15, 2022

RoMFAC: A robust mean-field actor-critic reinforcement learning against adversarial perturbations on states
Ziyuan Zhou, Guanjun Liu
Multi Agent Adversarial Perturbation Policy Gradient Multi Agent Deep Reinforcement Learning Actor Critic Reinforcement Learning State Specific Decision Making Mean Field Reinforcement Learning Adversarial Team Game

February 13, 2022

Individual-Level Inverse Reinforcement Learning for Mean Field Games
Yang Chen, Libo Zhang, Jiamou Liu, Shuyue Hu
Reward Function Inverse Reinforcement Learning Collective Behavior Mean Field Game Large Scale Multi Agent Mean Field Reinforcement Learning

December 16, 2021

Decentralized Mean Field Games
Sriram Ganapathi Subramanian, Matthew E. Taylor, Mark Crowley, Pascal Poupart
Mean Field Mean Field Game Mean Field Reinforcement Learning