Game-Theoretic Deep Reinforcement Learning to Minimize Carbon Emissions and Energy Costs for AI Inference Workloads in Geo-Distributed Data Centers.
Ninad HogadeSudeep PasrichaPublished in: CoRR (2024)
Keyphrases
- carbon dioxide
- game theoretic
- data center
- reinforcement learning
- game theory
- energy consumption
- decision problems
- greenhouse gas emissions
- energy efficiency
- cloud computing
- multi agent
- power generation
- energy saving
- nash equilibrium
- power plant
- cost effective
- climate change
- energy aware
- power consumption
- cooperative
- regret minimization
- nash equilibria
- database systems
- machine learning
- trust model
- energy supply
- optimal policy
- learning algorithm
- greenhouse gases
- electrical energy
- multi tenant
- bayesian networks
- imperfect information
- mobile agents
- minority game
- production cost
- multi agent systems
- special case
- state space