Abstract: Round Robin (RR), CPU scheduling algorithm, is widely accepted scheduling strategy for many time shared operating systems. The traditional Fixed quantum Round Robin (FRR) scheme works well for fair share scheduling though the large number of context switches lead to excessive system overhead. On the other hand, the available dynamic quantum based RR algorithm (DRR) reduces context switches though at the cost of fairness and results in service degradation to an individual process. In this paper, a new variant of RR named as adaptive Round Robin is proposed that trades off between fairness and context switches parameters in a judicious way and provide a better balancing among these conflicting parameters. The proposed algorithm chooses the time quantum adaptively based on existing burst time of the available jobs in the job pool. In this way, both smaller and larger size jobs get fair time for their execution. The performance of the suggested technique has been analyzed using extensive simulations on a wide variety of jobs. The paper also presents the comparative analysis of proposed algorithm with existing FRR and DRR scheduling algorithms on the basis of varying time quantum, average waiting time, average turnaround time, performance ratio and number of context switches.
Keywords: CPU scheduling, Round Robin CPU scheduling algorithm, Turnaround time, Waiting time, Context switching, Performance ratio, Simulation analysis.