Data-driven partially observable dynamic processes using adaptive dynamic programming.
Xiangnan ZhongZhen NiYufei TangHaibo HePublished in: ADPRL (2014)
Keyphrases
- data driven
- partially observable
- dynamic programming
- markov decision processes
- state space
- infinite horizon
- markov decision problems
- reinforcement learning
- partial observability
- dynamical systems
- decision problems
- dec pomdps
- dynamic environments
- belief state
- linear programming
- partially observable markov decision processes
- finite state
- stereo matching
- partially observable environments
- reward function
- optimal control
- learning algorithm
- orders of magnitude
- general purpose
- special case
- knowledge base