Policy-Gradient-Based Reinforcement Learning for Computing Resources Allocation in O-RAN.
Mahdi ShararaTurgay PamukluSahar HoteitVéronique VèqueMelike Erol-KantarciPublished in: CloudNet (2022)
Keyphrases
- computing resources
- reinforcement learning
- optimal policy
- allocation policy
- policy search
- cloud computing
- resource management
- model free
- limited resources
- markov decision process
- action selection
- resource allocation
- allocation policies
- markov decision processes
- reward function
- allocation strategy
- geographically distributed
- actor critic
- partially observable
- policy gradient
- control policy
- grid computing
- markov decision problems
- function approximation
- function approximators
- policy evaluation
- policy iteration
- state space
- virtual machine
- reinforcement learning algorithms
- action space
- temporal difference
- learning algorithm
- network resources
- high performance computing
- average reward
- load balance
- partially observable markov decision processes
- dynamic programming
- data processing
- multi core processors
- data management
- management system
- data center
- cooperative
- database
- agent learns
- data mining
- databases