Optimal Decision-Making in Mixed-Agent Partially Observable Stochastic Environments via Reinforcement Learning.
Roi CerenPublished in: CoRR (2019)
Keyphrases
- partially observable
- reinforcement learning
- decision making
- state space
- markov decision processes
- action selection
- decision problems
- partial observations
- initially unknown
- partial observability
- control policies
- dynamical systems
- partially observable domains
- dynamic programming
- reward function
- markov decision problems
- fully observable
- belief state
- partially observable environments
- infinite horizon
- action models
- multi agent
- function approximation
- optimal policy
- model free
- control policy
- optimal control
- decision makers
- multi agent systems
- hidden state
- average reward
- temporal difference
- data mining
- optimal solution
- partially observable markov decision process
- machine learning
- learning algorithm
- continuous state
- dynamic environments
- partially observable markov decision processes
- average cost
- transfer learning
- heuristic search
- domain specific
- mobile robot
- action space
- learning agent
- np hard
- reinforcement learning algorithms
- multiple agents