The preemptive EDF optimization based on DNA-Genetic Algorithm

Abstract

Earliest Deadline First (EDF) is a favorable algorithm in the field of dynamic scheduling which plans the task scheduling during system running. The CPU utilization of EDF algorithm may achieve to one hundred percent. Although EDF has many excellent performances and is used broadly in actual embedded real-time systems, it has many shortages. In order to reduce the release time of EDF scheduling algorithm and to decrease scheduling spending, the Deoxyribonucleic acid Genetic Algorithm (DNA-GA) is used to optimize the preemptive EDF scheduling algorithm in off-line model. Simulation experiments are carried out to compare the performances. Compared with the scheduling spending before optimized, the scheduling spending after optimized was reduced dramatically and the real-time performance was improved.