From preemptive to non-preemptive speed-scaling scheduling

Abstract : We are given a set of jobs, each one specified by its release date, its deadline and its processing volume (work), and a single (or a set of) speed-scalable processor(s). We adopt the standard model in speed-scaling in which if a processor runs at speed . s then the energy consumption is . sα units of energy per time unit, where . α>1 is a small constant. Our goal is to find a schedule respecting the release dates and the deadlines of the jobs so that the total energy consumption to be minimized. While most previous works have studied the preemptive case of the problem, where a job may be interrupted and resumed later, we focus on the non-preemptive case where once a job starts its execution, it has to continue until its completion without any interruption. As the preemptive case is known to be polynomially solvable for both the single-processor and the multiprocessor case, we explore the idea of transforming an optimal preemptive schedule to a non-preemptive one. We prove that the preemptive optimal solution does not preserve enough of the structure of the non-preemptive optimal solution, and more precisely that the ratio between the energy consumption of an optimal non-preemptive schedule and the energy consumption of an optimal preemptive schedule can be very large even for the single-processor case. Then, we focus on some interesting families of instances: (i) equal-work jobs on a single-processor, and (ii) agreeable instances in the multiprocessor case. In both cases, we propose constant factor approximation algorithms. In the latter case, our algorithm improves the best known algorithm of the literature. Finally, we propose a (non-constant factor) approximation algorithm for general instances in the multiprocessor case.
Type de document :
Article dans une revue
Discrete Applied Mathematics, Elsevier, 2015, 181, pp.11-20. 〈10.1016/j.dam.2014.10.007〉
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01082393
Contributeur : Frédéric Davesne <>
Soumis le : jeudi 13 novembre 2014 - 14:01:37
Dernière modification le : mercredi 21 mars 2018 - 18:58:11

Identifiants

Citation

Evripidis Bampis, Alexander Kononov, Dimitrios Letsios, Giorgio Lucarelli, Ioannis Nemparis. From preemptive to non-preemptive speed-scaling scheduling. Discrete Applied Mathematics, Elsevier, 2015, 181, pp.11-20. 〈10.1016/j.dam.2014.10.007〉. 〈hal-01082393〉

Partager

Métriques

Consultations de la notice

230