Quantile-Based Policy Optimization for Reinforcement Learning.
Jinyang JiangYijie PengJiaqiao HuPublished in: WSC (2022)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- global optimization
- state and action spaces
- approximate dynamic programming
- optimization algorithm
- markov decision process
- optimization method
- action selection
- actor critic
- transition model
- control policies
- control policy
- action space
- policy iteration
- markov decision problems
- partially observable environments
- partially observable
- constrained optimization
- function approximation
- markov decision processes
- optimization problems
- control problems
- reinforcement learning algorithms
- reward function
- function approximators
- optimization methods
- policy evaluation
- supervised learning
- least squares
- state space
- neural network