基于进程执行时间的多级反馈队列调度算法

MLFQ Scheduling Algorithm Based on Burst Time of the Now Running Process

针对多级反馈队列(MLFQ)调度算法在时间片大小选取上对系统性能的关键影响,提出了一种基于进程执行时间的多级反馈队列调度算法。算法结合动态时间量子思想,通过进程的执行时间动态确定队列以及时间片来完成调度。同时为了减少队列进程的切换次数,利用中位数的方法来决定时间片的大小。实验结果表明,与传统的多级反馈队列调度算法相比,改进的算法不仅缩短了进程的平均周转时间和平均等待时间,也减少了进程切换次数,为操作系统领域处理机调度智能化提供了有效的参考价值。

The time quantum in Multi-level feedback queue has a great influence on the performance of operating system.To solve the problem,a self-adjustment time quantum in MLFQ was proposed based on burst time of the now running process.The idea of dynamic time quantum was combined with,and the queues and time quantum were dynamicly adjusted according to the now running process.To reduce the switch times of the processes,the median was used to determine the value of time quantum.The experimental results show that the proposed algorithm has lower average turnaround time and average waiting time than other traditional MLFQ and a certain reduction in the switch times of processes.It is valuable for intelligent scheduling algorithm in the field of operating system.