Abstract
The Linux kernel's performance can be improved by customizing one or more kernel settings, or adjusting kernel parameters. A system administrator performs this work in order to ensure that systems are performing optimally. A desirable goal is to allow end users to experience the same level of performance enhancement using a self-tuning kernel. This research aims to address the problem of manually adjusting kernel parameters to affect performance as workload changes. This project describes a tuning system that utilizes genetic algorithms to tune kernel parameters to an objective performance goal. The purpose of this approach is for the genetic algorithm to find kernel parameters which affect system performance. This approach facilitates performance optimization for a range of different load types.