PI-ELM: Reinforcement learning-based adaptable policy improvement for dynamical system.
Yingbai HuXu WangYueyue LiuWeiping DingAlois KnollPublished in: Inf. Sci. (2023)
Keyphrases
- dynamical systems
- partially observable
- reinforcement learning
- state space
- optimal policy
- reinforcement learning methods
- partially observable markov decision processes
- reinforcement learning problems
- policy search
- differential equations
- markov decision problems
- hidden state
- markov decision process
- phase space
- dynamic systems
- action selection
- nonlinear dynamical systems
- reward function
- continuous state
- dynamical behavior
- action space
- infinite horizon
- policy gradient
- markov decision processes
- function approximation
- extreme learning machine
- neural network
- predictive state representations
- policy iteration
- policy evaluation
- reinforcement learning algorithms
- average reward
- dynamic programming
- immune network