Abstract:
The operating system (OS) acts as a resource manager whose responsibility is to manage the resources
of a computer system. Among all resources, the CPU is one of the most crucial resources that manage the
processes. Process management is achieved by a specific type of algorithm called CPU scheduling algorithm.
CPU scheduling is a vital task of the OS and the whole performance of the system depends on the CPU
scheduling criteria such as reducing waiting time, turnaround time, response time, and number of context
switches, while enhancing the CPU utilization. Several well-known CPU scheduling algorithms like First Come-First-Served (FCFS), Shortest Job First (SJF), Shortest Remaining Time First (SRTF), Highest
Response Ratio Next (HRRN), Round Robin (RR), and Priority Scheduling algorithms are coming into the
picture. In time-shared environment, RR CPU scheduling is preferred, but the system's performance depends
on choosing the most appropriate time quantum. By fusing the advantages of RR with features of SJF, this
paper provides an intuitive approach to enhance the conventional RR CPU scheduling algorithm with adaptive
time quanta that intends to improve system performance over the improver version like IRRVQ, MRR, Tajwar
et al. and Fiad et al. The work offers experimental evidence that the proposed algorithm Dynamic Round
Robin with Adaptive Time Quanta (DRRATQ) performs better than the conventional RR and other existing
work, by decreasing the waiting time, turnaround time, response time, and number of context switches.
Implementing this algorithm to a time-sharing or distributed environment will undoubtedly improve system
performance and help avoid issues like thrashing, incorporate aging, CPU affinity, and starvation. Since the
proposed scheduling is work-conserving in nature, it can be used for statistical multiplexing and best-effort
packet switching in a network packet scheduling environment.