An Optimised CPU Scheduling Algorithm With Adaptive Time Quantum Approach

Abstract

CPU scheduling is an essential mechanism implemented by the operating system to determine the execution of multiple processes by the CPU. The primary objective of the scheduling algorithms is to optimize the systems’ performance efficiently. The performance of a CPU scheduling algorithm depends on various factors and can be evaluated on various criteria like average turnaround time, average waiting time, throughput, fairness etc. This paper aims to present an optimal CPU scheduling algorithm, Adaptive Quantum Round Robin (AQRR) in a uniprocessor environment. Related work has been done on increasing the performance of existing Round Robin Algorithms using dynamic time quantum approaches, but the maximum percentage gain in turnaround time and waiting time using these approaches, over the traditional Round Robin algorithm is not more than 30% and 40% respectively. The proposed algorithm in this paper outperforms these algorithms in both turnaround time and waiting time. The AQRR algorithm is based on the standard Round Robin algorithm but it is integrated with a dynamic time quantum which is self-adaptive triggered on the event of the arrival of new processes in the ready queue or the completion of a process in the process queue. The dynamic time quantum is updated based on the remaining burst time of the running processes in the process queue at that instance as well as a weightage associated with it. Finally, a comparative study between the prevailing similar scheduling algorithms and the proposed algorithm is observed and noted, based on different scenarios. The following algorithms are compared on two criterias: average turnaround time and average waiting time.