Safe Deep Reinforcement Learning for Multi-Agent Systems with Continuous Action Spaces.
Ziyad SheebaelhamdKonstantinos ZisisAthina NisiotiDimitris GkouletsosDario PavlloJonas KöhlerPublished in: CoRR (2021)
Keyphrases
- action space
- reinforcement learning
- multi agent systems
- single agent
- state space
- markov decision processes
- state and action spaces
- multi agent
- real valued
- continuous state
- continuous state spaces
- state action
- action selection
- stochastic processes
- continuous action
- function approximators
- markov decision problems
- multi agent reinforcement learning
- multiple agents
- markov decision process
- control policies
- reinforcement learning methods
- finite state
- reinforcement learning algorithms
- control problems
- function approximation
- dynamic environments
- markov chain
- machine learning
- temporal difference
- model free
- policy search
- cooperative
- learning algorithm