Learning-Augmented Approximation for Unrelated-Machines Makespan Scheduling
Researchers have developed a new learning-augmented algorithm for makespan minimization on unrelated machines, a problem denoted as R||Cmax. This approach extends a framework previously used for selection problems to scheduling, aiming to improve approximation ratios by incorporating predictions of job assignments. The algorithm achieves a (1+ε)-approximation for accurate predictions, with the approximation degrading to a 2-approximation as prediction error increases. AI
