Split Q Learning: Reinforcement Learning with Two-Stream Rewards.
Baihan LinDjallel BouneffoufGuillermo A. CecchiPublished in: IJCAI (2019)
Keyphrases
- reinforcement learning
- function approximation
- reinforcement learning algorithms
- model free
- state space
- markov decision processes
- data streams
- action selection
- reward function
- optimal policy
- temporal difference
- eligibility traces
- reinforcement learning methods
- control problems
- partially observable
- reward shaping
- continuous state and action spaces
- multi agent
- function approximators
- machine learning
- state action space
- transfer learning
- sliding window
- dynamic programming
- learning algorithm
- optimal control
- temporal difference learning
- learning process
- cooperative
- learning agent
- state abstraction
- policy search
- exploration strategy