Heuristics for generalized task system
Document Type
Conference Proceeding
Publication Date
12-1-2003
Abstract
With the advance of parallel processors, Generalized Task System (GTS) has been proposed to capture the notion of parallelism within a single task Based on this model, the complexity of minimizing the total (or average) flaw time for the preemptive and non-preemptive scheduling discipline has been proved to be binary NP-hard even when there are only two processors and each task has exactly two independent subtasks. In this paper, we propose a preemptive approximation algorithm to minimize the total flow time and its worst-case and average performance analyzed.
Identifier
1642282119 (Scopus)
ISBN
[1892512416]
Publication Title
Proceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications
First Page
1447
Last Page
1453
Volume
3
Recommended Citation
Ng, Chung W.; Leung, Joseph Y.T.; and Young, G., "Heuristics for generalized task system" (2003). Faculty Publications. 13890.
https://digitalcommons.njit.edu/fac_pubs/13890
