A continuous time markov decision process based on-chip buffer allocation methodology.
Sankalp KallakuriNattawut ThepayasuwanAlex DoboliEugene A. FeinbergPublished in: ACM Great Lakes Symposium on VLSI (2005)
Keyphrases
- markov decision process
- state space
- buffer allocation
- markov decision processes
- infinite horizon
- reinforcement learning
- optimal policy
- stationary policies
- production line
- markov chain
- optimal control
- queueing networks
- initial state
- low cost
- transition probabilities
- dynamic programming
- dynamical systems
- buffer management
- finite state
- reward function
- linear programming
- search space