Model-Free Learning and Optimal Policy Design in Multi-Agent MDPs Under Probabilistic Agent Dropout.
Carmel FisckoSoummya KarBruno SinopoliPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- optimal policy
- model free
- multi agent
- markov decision processes
- policy iteration
- state space
- average reward
- markov decision process
- learning algorithm
- reward function
- function approximation
- partially observable
- learning agent
- reinforcement learning algorithms
- finite horizon
- reinforcement learning methods
- temporal difference
- action selection
- infinite horizon
- bayesian reinforcement learning
- policy evaluation
- multi agent systems
- supervised learning
- markov decision problems
- decision problems
- finite state
- multistage
- multiple agents
- rl algorithms
- learning problems
- function approximators
- dynamic programming
- temporal difference learning
- machine learning
- stochastic games
- optimal control
- decision theoretic
- initial state
- single agent
- state dependent