A Comparative Analysis of Quantum Time based on Arithmetic, Geometric, and Harmonic Mean for Dynamic Round Robin Scheduling

Abstract

Round Robin Scheduling (RRS) is a job-scheduling algorithm considered to be fair and cycle executive as it handles requests in a circular first-in-first-out (FIFO) order. This same characteristic emphasizes its starvation free nature, achieved by using proper quantum time. In this paper, we compare and analyze different means of computing for the quantum time in each cycle of the dynamic RRS. The first method is Arithmetic Mean which is simply taking the sum of all process burst times and then dividing it by the total number of processes. The second method is Geometric Mean computed by adding the burst times in the series then taking its square root. The last method is Harmonic Mean which is calculated by dividing the number of processes by the reciprocal of each burst time in the series. These methods provided a better scheduling criterion in measuring the average turnaround time, average waiting time and the total number of context switches in RRS. Results show that Harmonic Mean proves to be more efficient in computing the quantum time for RRS CPU scheduling as it gives equal weight to each data point. Evidenced by the outcome of varied test cases, the results remain steady due to the decrease in average turnaround time, average waiting time and the total number of context switches. Succeeding works can implement it on hard real-time systems, alter the geometric mean formula, illustrate and triangulate the computed quantum time and experiment on broader test cases.