A Hybrid Adaptive CPU Scheduling Algorithm Using a Dynamic Time Quantum and Simple Additive Weighting

Abstract

CPU scheduling is a fundamental function in operating systems that determines process execution order to optimize system performance. Classical algorithms such as FCFS, SJF, Priority, and Round Robin often focus on limited criteria, leading to issues such as unfairness, starvation, and excessive context-switch overhead. This paper proposes a Hybrid Adaptive CPU Scheduling Algorithm based on Dynamic Time Quantum and Simple Additive Weighting (HACS-DTQSAW). The proposed approach integrates multi-criteria decision-making with adaptive time-sharing by evaluating processes using the Simple Additive Weighting (SAW) method across multiple attributes, including arrival time, remaining burst time, and priority. The time quantum is dynamically computed using both the mean and median of remaining burst times to ensure balanced scheduling behavior. The proposed algorithm is evaluated using extensive simulation across diverse workloads and compared with classical and modern scheduling algorithms. Experimental results demonstrate significant improvements in efficiency, fairness, and starvation prevention, with statistically significant performance gains, p < 0.001. The results indicate that the proposed approach provides a robust and scalable solution for modern CPU scheduling environments.