Promoting Coordination through Policy Regularization in Multi-Agent Deep Reinforcement Learning.
Julien RoyPaul BardeFélix G. HarveyDerek NowrouzezahraiChris PalPublished in: NeurIPS (2020)
Keyphrases
- multi agent
- reinforcement learning
- optimal policy
- multi agent reinforcement learning
- cooperative multi agent systems
- policy search
- action selection
- multi agent systems
- multiagent systems
- cooperative
- multiagent learning
- single agent
- multi agent coordination
- markov decision process
- multiple agents
- state space
- partially observable markov decision processes
- actor critic
- reward function
- policy gradient
- control policies
- reinforcement learning algorithms
- approximate dynamic programming
- reinforcement learning problems
- function approximation
- action space
- partially observable
- markov decision problems
- heterogeneous agents
- control policy
- function approximators
- markov decision processes
- policy iteration
- model free
- average reward
- learning agents
- multi agent environments
- partially observable environments
- state action
- multi agent planning
- temporal difference
- autonomous agents
- software agents
- policy evaluation
- rl algorithms
- agent communication
- continuous state spaces
- state and action spaces
- reinforcement learning agents
- transfer learning
- infinite horizon
- intelligent agents
- decision problems
- learning problems
- partially observable markov decision process
- inverse reinforcement learning
- machine learning
- partially observable domains
- policy gradient methods
- regularization parameter
- learning algorithm
- agent learns
- long run
- management policies
- partial observability
- optimal control
- continuous state