Two-Stage Reinforcement Learning Policy Search for Grid-Interactive Building Control.
Xiangyu ZhangYue ChenAndrey BernsteinRohit ChintalaPeter A. GrafXin JinDavid BiagioniPublished in: IEEE Trans. Smart Grid (2022)
Keyphrases
- policy search
- reinforcement learning
- control problems
- reinforcement learning algorithms
- continuous state
- optimal control
- control policy
- control system
- control policies
- markov decision processes
- reward function
- learning algorithm
- continuous action
- control strategies
- function approximation
- dynamic programming
- function approximators
- policy gradient
- multi agent
- model free
- action selection
- finite state
- control strategy
- transfer learning
- markov decision problems
- multi agent systems
- machine learning