A carbon-aware Spark scheduler that defers low-priority tasks during high-carbon periods, using importance scores from an ML scheduler, reduced carbon by roughly a third in a 100-node prototype with near-neutral end-to-end completion time.
Near-Optimal Consistency-Robustness Trade-Offs for Learning-Augmented Online Knapsack Problems
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
This paper introduces a family of learning-augmented algorithms for online knapsack problems that achieve near Pareto-optimal consistency-robustness trade-offs through a simple combination of trusted learning-augmented and worst-case algorithms. Our approach relies on succinct, practical predictions -- single values or intervals estimating the minimum value of any item in an offline solution. Additionally, we propose a novel fractional-to-integral conversion procedure, offering new insights for online algorithm design.
citation-role summary
background 1
citation-polarity summary
fields
cs.DC 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Carbon- and Precedence-Aware Scheduling for Data Processing Clusters
A carbon-aware Spark scheduler that defers low-priority tasks during high-carbon periods, using importance scores from an ML scheduler, reduced carbon by roughly a third in a 100-node prototype with near-neutral end-to-end completion time.