Kernel Taylor-Based Value Function Approximation for Continuous-State Markov Decision Processes.
Junhong XuKai YinLantao LiuPublished in: CoRR (2020)
Keyphrases
- markov decision processes
- continuous state
- action space
- state action
- finite state
- state space
- reinforcement learning
- partially observable markov decision processes
- stochastic games
- policy iteration
- average reward
- markov decision process
- optimal policy
- temporal difference
- continuous state spaces
- dynamic programming
- kernel methods
- reward function
- kernel function
- approximate dynamic programming
- planning problems
- reinforcement learning algorithms
- average cost
- partially observable
- basis functions
- infinite horizon
- control policies
- markov chain
- function approximators
- function approximation
- decision problems
- finite horizon
- model free
- heuristic search
- initial state
- real valued
- learning algorithm
- policy gradient
- long run
- machine learning