DocumentCode :
263696
Title :
A DAG Task Scheduling Scheme on Heterogeneous Computing Systems Using Invasive Weed Optimization Algorithm
Author :
Kenli Li ; Shuai Li ; Yuming Xu ; Zhaoxin Xie
Author_Institution :
Coll. of Inf. Sci. & Eng., Hunan Univ., Changsha, China
fYear :
2014
fDate :
13-15 July 2014
Firstpage :
262
Lastpage :
267
Abstract :
Efficient task scheduling is crucial to heterogeneous cluster performance. And various scheduling methods based on random search technique have been proposed for both homogeneous and heterogeneous cluster systems. However, most of these methods have high computational overhead and poor convergence. Invasive weed optimization algorithm (IWO) is a novel bionic intelligent optimization algorithm that has fast convergence rate and easier implementation than traditional genetic algorithm (GA) based algorithm. In this paper, an IWO task scheduling (IWOTS) algorithm is proposed for heterogeneous cluster system. To the best of our knowledge, this study is the first time to apply IWO to discrete task scheduling problems. Extensive simulation experiment results show that IWOTS generally exhibits outstanding convergence performance and could produce an optimal scheduling strategy.
Keywords :
computational complexity; optimisation; processor scheduling; scheduling; search problems; DAG task scheduling scheme; IWOTS; NP-complete problem; bionic intelligent optimization algorithm; discrete task scheduling problem; heterogeneous cluster system; heterogeneous computing system; invasive weed optimization algorithm; multiprocessor scheduling; random search technique; Algorithm design and analysis; Clustering algorithms; Optimal scheduling; Processor scheduling; Program processors; Scheduling; DAG scheduling; Invasive weed optimization; genetic algorithms; heterogeneous systems; task graphs;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Parallel Architectures, Algorithms and Programming (PAAP), 2014 Sixth International Symposium on
Conference_Location :
Beijing
ISSN :
2168-3034
Print_ISBN :
978-1-4799-3844-5
Type :
conf
DOI :
10.1109/PAAP.2014.34
Filename :
6916475
Link To Document :
بازگشت