Rapid Convergecast on Commodity Hardware: Performance Limits and Optimal Policies.
Haibo ZhangFredrik ÖsterlindPablo SoldatiThiemo VoigtMikael JohanssonPublished in: SECON (2010)
Keyphrases
- optimal policy
- commodity hardware
- markov decision processes
- decision problems
- parallel computing
- parallel processing
- cloud computing
- state space
- finite horizon
- dynamic programming
- infinite horizon
- reinforcement learning
- long run
- sufficient conditions
- multistage
- data analytics
- shared memory
- big data
- average reward
- massively parallel
- computing systems
- high end
- parallel execution
- serial inventory systems
- markov chain
- graphical models
- average cost
- open source
- map reduce
- general purpose
- knowledge discovery
- probabilistic model
- average reward reinforcement learning