Accelerating Reinforcement Learning via Predictive Policy Transfer in 6G RAN Slicing.
Ahmad M. NagibHatem Abou-ZeidHossam S. HassaneinPublished in: IEEE Trans. Netw. Serv. Manag. (2023)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- transfer learning
- markov decision process
- action selection
- reinforcement learning algorithms
- state space
- control policies
- policy evaluation
- partially observable environments
- action space
- reinforcement learning problems
- reward function
- state and action spaces
- actor critic
- control policy
- function approximators
- policy gradient
- markov decision problems
- function approximation
- state action
- machine learning
- rl algorithms
- model free
- infinite horizon
- markov decision processes
- learning tasks
- approximate dynamic programming
- learning algorithm
- dynamic programming
- decision problems
- knowledge transfer
- cross domain
- optimal control
- temporal difference
- finite state
- policy iteration
- transition model
- partially observable markov decision processes
- continuous state spaces
- average reward
- learning classifier systems
- active learning
- multi agent
- neural network
- allocation policy
- transferring knowledge
- continuous state
- previously learned
- temporal difference learning
- state dependent
- real robot
- control problems
- least squares