Dynamic Choice of State Abstraction in Q-Learning.
Marco TamassiaFabio ZambettaWilliam L. RaffeFlorian 'Floyd' MuellerXiaodong LiPublished in: ECAI (2016)
Keyphrases
- state abstraction
- state space
- reinforcement learning
- hierarchical reinforcement learning
- variable resolution
- dynamic programming
- function approximation
- learning algorithm
- multi agent
- optimal policy
- markov decision processes
- reinforcement learning algorithms
- path finding
- mobile robot
- dynamic environments
- model free