Shield Decentralization for Safe Reinforcement Learning in General Partially Observable Multi-Agent Environments.
Daniel MelcerChristopher AmatoStavros TripakisPublished in: AAMAS (2024)
Keyphrases
- partially observable
- reinforcement learning
- multi agent environments
- state space
- markov decision processes
- partial observability
- partially observable environments
- partially observable domains
- decision problems
- function approximation
- dynamical systems
- reinforcement learning algorithms
- special case
- reinforcement learning methods
- markov decision problems
- multi agent
- control problems
- belief state
- multi agent systems
- infinite horizon
- reward function
- learning algorithm
- knowledge base
- model free
- autonomous agents
- probability distribution
- machine learning
- transfer learning
- decision making
- heuristic search