Risk-Sensitive Bayesian Games for Multi-Agent Reinforcement Learning under Policy Uncertainty.
Hannes ErikssonDebabrota BasuMina AlibeigiChristos DimitrakakisPublished in: CoRR (2022)
Keyphrases
- multi agent reinforcement learning
- risk sensitive
- stochastic games
- markov decision processes
- learning agents
- decision theory
- expected utility
- optimal policy
- average reward
- markov decision problems
- control policies
- reinforcement learning
- utility function
- game theory
- optimality criterion
- reward function
- model free
- optimal control
- multi agent learning
- policy iteration
- infinite horizon
- average cost
- partially observable
- nash equilibria
- finite state
- finite horizon
- markov decision process
- efficient optimization
- robust optimization
- decision theoretic
- multi agent
- decision problems
- state space
- artificial intelligence
- reinforcement learning algorithms
- nash equilibrium
- multi agent systems
- action space
- control policy
- learning agent
- incomplete information
- long run
- pareto optimal
- bayesian networks
- decision makers
- decision making
- imperfect information
- dynamic programming
- control system
- learning algorithm