Reinforcement Learning in Reward-Mixing MDPs.
Jeongyeol KwonYonathan EfroniConstantine CaramanisShie MannorPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- markov decision processes
- reward function
- function approximation
- state space
- reinforcement learning algorithms
- average reward
- optimal policy
- machine learning
- continuous state and action spaces
- eligibility traces
- model free
- temporal difference
- action sets
- learning algorithm
- control problems
- dynamic programming
- policy search
- policy evaluation
- markov decision problems
- action selection
- multi agent
- reinforcement learning methods
- total reward
- action space
- policy iteration
- markov decision process
- model based reinforcement learning
- partially observable
- learning capabilities
- reward shaping
- search space
- planning under uncertainty
- continuous state
- linear programming
- function approximators
- utility function
- fixed point