Robust Non-Clairvoyant Scheduling with Classification Models
A confusion-matrix robustness framework yields optimal non-adaptive scheduling when job times are unknown until completion.
The paper studies non-clairvoyant single-machine scheduling to minimize the sum of completion times, where processing times stay unknown until jobs finish. It uses a classifier confusion matrix to treat uncertainty as permutations within predicted classes rather than numerical scenarios. An optimal non-adaptive strategy is oblivious to Min-Max, Min-Max Regret, and expected-objective criteria. Adaptive and randomized algorithms can outperform that strategy when the confusion matrix has particular structure.
- Jobs are grouped by a classifier with a known confusion matrix.
- Uncertainty is modeled as permutations inside predicted classes.
- One non-adaptive policy is optimal for three robust criteria.
- Adaptive and randomized methods can improve on structured matrices.
From the original article
1 min read at arxiv.orgWe study the classical single-machine scheduling problem of minimizing the sum of completion times of jobs in a non-clairvoyant setting, where the processing time of each job remains unknown until its completion. This is a hard problem for which no constant competitive algorithm is possible.
By Anthony Dugois · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.