JCOM_V2_N1_RP1
Comparison of some Algorithms for Multi-Processor Job Scheduling Problem based on the Random Nature of Job Completion Times
Ruwanthini Siyambalapitiya
Manjula Sandirigama
Journal on Computer Science
2347–6141
2
1
6
12
Makespan, Lower Bound, Approximation Algorithms, Problem Size Ratio
In this paper, the authors wish to report some further computational results related to two algorithms proposed earlier for the multi-processor job scheduling problem. Here, they have compared the performance of an FCFS-based algorithm for multi-processor scheduling with a greedy-based algorithm known as Decreasing-Ascend algorithm. They have considered the random nature of job completion times, to get a deeper insight into the performance of the algorithms. More than 20,000 data sets were created with varying combinations of jobs with shorter job lengths and longer job lengths. They keep the total execution time (sum of individual job durations) as fixed for all the instances considered and have showed that even if we consider this random situation, the performance level of the algorithms reported earlier is still applicable.
March - May 2014
Copyright © 2014 i-manager publications. All rights reserved.
i-manager Publications
http://www.imanagerpublications.com/Article.aspx?ArticleId=2845