Model-Based RL for Mean-Field Games is not Statistically Harder than Single-Agent RL.
Jiawei HuangNiao HeAndreas KrausePublished in: CoRR (2024)
Keyphrases
- single agent
- learning agents
- reinforcement learning
- multi agent
- action space
- model free
- multiple agents
- stochastic games
- dynamic environments
- decision problems
- multi agent systems
- policy gradient
- markov decision processes
- learning agent
- exploration strategy
- optimal policy
- function approximation
- multi agent coordination
- reinforcement learning algorithms
- window search
- multiagent learning
- state space
- action selection
- game theory
- markov decision process
- np hard
- dec pomdps
- learning classifier systems
- temporal difference
- learning algorithm
- game playing
- path planning
- dynamic programming
- cooperative
- decision making