How good are SPT schedules for fair optimality criteria

Eric Angel (1) , Evripidis Bampis (1) , Fanny Pascual (2)

1 IBISC - Informatique, Biologie Intégrative et Systèmes Complexes
2 RO - Recherche Opérationnelle

We consider the following scheduling setting: a set of n tasks have to be executed on a set of m identical machines. It is well known that shortest processing time (SPT) schedules are optimal for the problem of minimizing the total sum of completion times of the tasks. In this paper, we measure the quality of SPT schedules, from an approximation point of view, with respect to the following optimality criteria: sum of completion times per machine, global fairness, and individual fairness.

Mots clés: Multiprocessor scheduling; SPT; Fairness measures; Approximation algorithms

Domaines: Informatique; Recherche opérationnelle [math.OC]; Algorithme et structure de données [cs.DS]; Mathématique discrète [cs.DM]; Informatique et théorie des jeux [cs.GT]

Eric Angel, Evripidis Bampis, Fanny Pascual. How good are SPT schedules for fair optimality criteria. Annals of Operations Research, 2008, 159 (1), pp.53--64. ⟨10.1007/s10479-007-0267-0⟩. ⟨hal-00341351⟩