Develops O(1)-competitive learning-augmented scheduling algorithms with O(1) preemptions per job for single and unrelated machines, with logarithmic overhead on prediction error, and first such guarantees for unrelated and malleable machines.
While a machinei could technically idle after meeting the specific processing targets vj,k for its assigned jobs, such idling results in an unnecessary loss of efficiency
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Learning-Augmented Online Scheduling with Parsimonious Preemption
Develops O(1)-competitive learning-augmented scheduling algorithms with O(1) preemptions per job for single and unrelated machines, with logarithmic overhead on prediction error, and first such guarantees for unrelated and malleable machines.