Improving generalization in reinforcement learning through forked agents.
Olivier MoulinVincent François-LavetMark HoogendoornPublished in: CoRR (2022)
Keyphrases
- reinforcement learning
- multi agent
- multi agent systems
- learning agents
- multiagent systems
- action selection
- multi agent reinforcement learning
- intelligent agents
- agent receives
- multiagent learning
- multiple agents
- multi agent environments
- autonomous agents
- software agents
- single agent
- learning agent
- cooperative
- robocup soccer
- reinforcement learning agents
- agent behavior
- learning capabilities
- agent architecture
- multiagent reinforcement learning
- model free
- resource allocation
- state space
- learning problems
- agent interactions
- partial observability
- function approximation
- learning algorithm
- distributed agents
- reinforcement learning algorithms
- dynamic environments
- markov decision processes
- decision making
- learning process
- decentralized control
- artificial agents
- learned knowledge
- complex domains
- reasoning process
- partially observable
- optimal policy
- game theoretic
- agent model
- coalition formation
- optimal control
- incomplete information