Incremental Reinforcement Learning in Continuous Spaces via Policy Relaxation and Importance Weighting.
Zhi WangHan-Xiong LiChunlin ChenPublished in: IEEE Trans. Neural Networks Learn. Syst. (2020)
Keyphrases
- reinforcement learning
- action space
- optimal policy
- continuous state spaces
- policy search
- state space
- action selection
- markov decision process
- markov decision processes
- control policies
- state and action spaces
- discrete data
- incremental learning
- policy evaluation
- continuous state
- reinforcement learning algorithms
- policy iteration
- function approximation
- continuous state and action spaces
- partially observable markov decision processes
- reinforcement learning problems
- function approximators
- control policy
- partially observable domains
- markov decision problems
- policy gradient
- reward function
- control problems
- actor critic
- temporal difference
- model free
- partially observable environments
- dynamic programming
- machine learning
- fitted q iteration
- state dependent
- continuous domains
- partially observable
- state action
- policy making
- single agent
- average cost
- long run
- real valued
- decision problems
- supervised learning
- objective function
- policy gradient methods
- learning algorithm