Ratio-Based Aggressive Reclaim Algorithm for Mixed Task Sets

Abstract

This paper aims to investigate the scheduling problem of a mixed task set in a real-time system. When a mixed task set is executed, a scheduling algorithm should consider system energy consumption and the average response time of aperiodic jobs. This paper proposes the Ratio-based Aggressive Reclaim Algorithm (RARA) that reclaims unused execution time left by earlier-completed jobs and possible slack time reserved by the Total Bandwidth (TB) server but not used by aperiodic jobs for next jobs. A user-defined ratio is used to flexibly control the amount of slack time. A series of experiments are conducted to evaluate the proposed algorithm. The experimental results demonstrate that the RARA scheme with a suitable value of the user-defined variable has a better performance compared with the Mutual Reclaiming Scheme and the Bandwidth Sharing Scheme. From the experimental results, we observe that a smaller ratio value will significantly improve the system energy consumption and the average response time of aperiodic jobs.