+44 (0)1509 263171
Please use this identifier to cite or link to this item:
|Title: ||Multiprocessor speed scaling for jobs with arbitrary sizes and deadlines|
|Authors: ||Bell, Paul C.|
Wong, Prudence W.H.
|Keywords: ||Online algorithms|
Dynamic speed scaling
|Issue Date: ||2015|
|Publisher: ||© Springer|
|Citation: ||BELL, P.C. and WONG, P.W.H., 2015. Multiprocessor speed scaling for jobs with arbitrary sizes and deadlines. Journal of Combinatorial Optimization, 29 (4), pp. 739 - 749|
|Abstract: ||In this paper we study energy efficient deadline scheduling on multiprocessors in which the processors consumes power at a rate of sα when running at speeds, where α ≥ 2. The problem is to dispatch jobs to processors and determine the speed and jobs to run for each processor so as to complete all jobs by their deadlines using the minimum energy. The problem has been well studied for the single processor case. For the multiprocessor setting, constant competitive online algorithms for special cases of unit size jobs or arbitrary size jobs with agreeable deadlines have been proposed by Albers et al. (2007). A randomized algorithm has been proposed for jobs of arbitrary sizes and arbitrary deadlines by Greiner et al. (2009). We propose a deterministic online algorithm for the general setting and show that it is O(logαP)-competitive, where P is the ratio of the maximum and minimum job size.|
|Description: ||This article was published in the journal, Journal of Combinatorial Optimization [© Springer]. The definitive version is available at: http://dx.doi.org/10.1007/s10878-013-9618-8.
A preliminary version appeared in Proceedings of the 8th Annual Conference on Theory and Applications of Models of Computation, 2011, pp. 27–36.|
|Sponsor: ||This work is partially supported by EPSRC Grant EP/E028276/1.|
|Version: ||Accepted for publication|
|Publisher Link: ||http://dx.doi.org/10.1007/s10878-013-9618-8|
|Appears in Collections:||Published Articles (Computer Science)|
Files associated with this item:
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.