Multigrid Reinforcement Learning with Reward Shaping.
Marek GrzesDaniel KudenkoPublished in: ICANN (1) (2008)
Keyphrases
- reward shaping
- reinforcement learning
- complex domains
- reinforcement learning algorithms
- markov decision problems
- state space
- optimal policy
- learning algorithm
- multi agent
- neural network
- model free
- function approximation
- temporal difference
- markov decision processes
- action selection
- reward function
- least squares
- markov decision process
- domain knowledge
- transition model
- machine learning