Transactions on Machine Learning and Data MiningP-ISSN: 1865-6781, E-ISSN 2509-9337ISBN 978-3-942952-89-7
Volume 14 - Number 2 - October 2021 - Pages 47-54
Machine learning methods for CPU scheduling
Majed AlsaneaArab East Colleges Arabeast Colleges, Riyadh, Saudi Arabia
Abstract
CPU scheduling is an important component of modern operating systems.
The performance of any operating system is highly dependent on the scheduling dis-cipline utilized.
There are relatively few scheduling disciplines still in use. Almost all modern operating systems utilize
a Multilevel Feedback Queue scheduling al-gorithm (MLFQ). However, a fixed time slice set for each queue
poses a perfor-mance-inhibiting problem that afflicts this discipline.
The assignment of fixed time slices negatively influences the performance of the scheduler.
In this paper, we try to tackle this issue and enhance the efficiency of the MLFQ scheduler
by applying SMOreg, Support Vector Machines and Random Tree classification methods to approximate the
time slice required for each queue. We apply these methods to three artificial datasets, and the performance results are noted and compared.
Keywords: Support Vector Machine, Machine Learning, Random Tree Classification.
Download Paper (379 KB)