An Automated VNF Manager based on Parameterized Action MDP and Reinforcement Learning.
Xinrui LiNancy SamaanAhmed KarmouchPublished in: ICC (2021)
Keyphrases
- reinforcement learning
- markov decision processes
- action space
- action sets
- optimal policy
- markov decision process
- state space
- action selection
- state action
- reward shaping
- partially observable domains
- semi automated
- state and action spaces
- partially observable
- reward function
- decision theoretic planning
- discounted reward
- initial state
- function approximation
- reinforcement learning algorithms
- markov decision problems
- continuous state
- transition model
- learning algorithm
- policy iteration
- model free
- dynamic programming
- average reward
- fitted q iteration
- reinforcement learning methods
- finite state
- transfer learning
- management system
- machine learning
- factored markov decision processes
- bayesian reinforcement learning
- multi agent
- search algorithm
- expected reward
- mobile robot
- heuristic search
- sensory inputs
- function approximators