Parameterized Deep Reinforcement Learning With Hybrid Action Space for Edge Task Offloading.
Ting WangYuxiang DengZhao YangYang WangHaibin CaiPublished in: IEEE Internet Things J. (2024)
Keyphrases
- action space
- reinforcement learning
- state space
- markov decision processes
- state and action spaces
- real valued
- continuous state
- stochastic processes
- control policies
- reinforcement learning methods
- continuous state spaces
- dynamic programming
- function approximation
- action selection
- learning algorithm
- markov decision process
- state action
- multi agent
- single agent
- function approximators
- transfer learning
- control problems
- reward function
- evaluation function
- optimal policy
- markov decision problems
- hidden markov models
- machine learning