Graph Reinforcement Learning-based CNN Inference Offloading in Dynamic Edge Computing.
Nan LiAlexandros IosifidisQi ZhangPublished in: CoRR (2022)
Keyphrases
- reinforcement learning
- weighted graph
- disjoint paths
- dynamic environments
- function approximation
- dynamic networks
- markov decision processes
- directed graph
- cellular neural networks
- directed acyclic graph
- graph model
- undirected graph
- neural network
- random walk
- clique tree
- reinforcement learning algorithms
- graph theoretic
- graph representation
- edge information
- state space
- graph structure
- graph theory
- bipartite graph
- probabilistic inference
- optimal policy