Concept Accessibility as Basis for Evolutionary Reinforcement Learning of Dots and Boxes.
Anthony KnittelTerry BossomaierAllan SnyderPublished in: CIG (2007)
Keyphrases
- reinforcement learning
- genetic algorithm
- mathematical theory
- state space
- database
- multi agent
- concept learning
- optimal policy
- starting point
- robotic control
- reinforcement learning algorithms
- function approximation
- evolutionary computation
- markov decision processes
- dynamic programming
- special case
- evolutionary algorithm
- artificial intelligence
- machine learning