Highly Constrained Task Scheduling on Multiprocessing Systems | ||||
| Journal of the ACS Advances in Computer Science | ||||
| Article 5, Volume 1, Issue 1, 2007, Page 51-61 PDF (779.01 K) | ||||
| Document Type: Original Article | ||||
| DOI: 10.21608/asc.2007.147561 | ||||
 View on SCiNiTO | ||||
| Abstract | ||||
| The problem of non-preemptively scheduling a set of m tasks on n processors with communication overhead subject to precedence, memory and deadline constraints is considered. A new heuristic with the time complexity of 0(m2n), Augmented Least Space-Time First (LSTF), is proposed to minimize the maximum tardiness. The efficiency of the augmented LSTF using a large number of randomly-generated graphs and three real-world structures is compared with that of the augmented Earliest Deadline First- Earliest Task First (EDF-E) that schedules each ready task on the processor at which it can be scheduled at the earliest time and with that of EDF-R that select the processor at random. The result of the comparisons, which are based on the maximum tardiness and the number of tasks that miss their deadlines, indicates that the augmented LSTF outperforms both EDF-E and EDF-R. | ||||
|
Statistics Article View: 80 PDF Download: 117 |
||||
