Towards Cost-Optimal Policies for DAGs to Utilize IaaS Clouds with Online Learning.
Xiaohu WuHan YuGiuliano CasaleGuanyu GaoPublished in: CoRR (2021)
Keyphrases
- optimal policy
- online learning
- cloud computing
- average cost
- markov decision processes
- lost sales
- infinite horizon
- decision problems
- finite horizon
- dynamic programming
- long run
- state space
- finite state
- reinforcement learning
- multistage
- dynamic programming algorithms
- average reward
- expected cost
- sufficient conditions
- e learning
- directed acyclic graph
- average reward reinforcement learning
- state dependent
- machine learning
- active learning
- demand distributions
- long run average cost
- finite number
- total cost
- markov decision process
- initial state
- control policies
- order quantity
- linear program
- learning algorithm
- policy iteration
- inventory control
- inventory models
- optimal control
- inventory policy
- semi markov decision processes
- least squares