Approximate dynamic programming using fluid and diffusion approximations with applications to power management.
Wei ChenDayu HuangAnkur A. KulkarniJayakrishnan UnnikrishnanQuanyan ZhuPrashant G. MehtaSean P. MeynAdam WiermanPublished in: CDC (2009)
Keyphrases
- power management
- approximate dynamic programming
- diffusion approximations
- heavy traffic
- queueing networks
- power consumption
- linear program
- stochastic dynamic programming
- dynamic programming
- energy saving
- reinforcement learning
- long run
- energy efficiency
- control policy
- energy consumption
- steady state
- data center
- average cost
- step size
- arrival rate
- asymptotically optimal
- policy iteration
- multi agent
- queue length
- markov decision processes
- energy efficient
- low cost
- wireless sensor networks
- machine learning