Partially decentralized reinforcement learning in finite, multi-agent Markov decision processes.
Omkar J. TilakSnehasis MukhopadhyayPublished in: AI Commun. (2011)
Keyphrases
- markov decision processes
- multi agent
- reinforcement learning
- state and action spaces
- optimal policy
- state space
- reinforcement learning algorithms
- action space
- decentralized control
- finite state
- cooperative
- transition matrices
- policy iteration
- dynamic programming
- action sets
- model free
- partially observable
- model based reinforcement learning
- multiagent systems
- multi agent systems
- average cost
- average reward
- function approximation
- finite horizon
- markov decision process
- stationary policies
- partially observable markov decision process
- reachability analysis
- planning under uncertainty
- dec pomdps
- decision theoretic planning
- single agent
- learning algorithm
- machine learning
- control problems
- temporal difference
- markov games
- multiple agents
- multiagent reinforcement learning
- partially observable markov decision processes
- approximate dynamic programming
- reward function
- infinite horizon
- markov decision problems
- stochastic games
- action selection
- policy gradient
- factored mdps
- continuous state
- optimal control
- decision problems