Customization of Arbitration Policies and Buffer Space Distribution Using Continuous-Time Markov Decision Processes.
Sankalp S. KallakuriAlex DoboliPublished in: IEEE Trans. Very Large Scale Integr. Syst. (2007)
Keyphrases
- markov decision processes
- optimal policy
- stationary policies
- state space
- markov decision process
- average cost
- policy iteration algorithm
- decision processes
- finite state
- dynamic programming
- reward function
- reinforcement learning
- discounted reward
- total reward
- decentralized control
- transition matrices
- finite horizon
- policy iteration
- reachability analysis
- partially observable markov decision processes
- action sets
- risk sensitive
- average reward
- macro actions
- action space
- optimal control
- dynamical systems
- decision problems
- long run
- markov chain
- planning under uncertainty
- infinite horizon
- control policies
- expected reward
- decision theoretic planning
- state and action spaces
- partially observable
- factored mdps
- multistage
- reinforcement learning algorithms
- hierarchical reinforcement learning
- sufficient conditions
- model based reinforcement learning
- transition probabilities
- real time dynamic programming
- machine learning