Adaptive State Space Partitioning of Markov Decision Processes for Elastic Resource Management.
Konstantinos LolosIoannis KonstantinouVerena KantereNectarios KozirisPublished in: ICDE (2017)
Keyphrases
- markov decision processes
- resource management
- state space
- optimal policy
- reinforcement learning
- resource allocation
- management system
- finite state
- policy iteration
- dynamic programming
- heuristic search
- reinforcement learning algorithms
- quality of service
- reachability analysis
- factored mdps
- decision theoretic planning
- partially observable
- dynamical systems
- action space
- planning under uncertainty
- intelligent agents
- markov chain
- grid computing
- transition matrices
- average cost
- action sets
- state and action spaces
- macro actions
- average reward
- model based reinforcement learning
- markov decision process
- state abstraction
- state variables
- java platform
- data mining
- real time dynamic programming
- markov decision problems
- decision processes
- linear programming
- search algorithm
- learning algorithm