Adaptive Scheduling Heuristic Priority Linear Regression (ASH-PLR): A Novel CPU Scheduling Algorithm using Predictive Priority Levels

Abstract

Inadequate implementation of parameters such as priority levels can be seen in many common and contemporary scheduling algorithms which can lead to starvation. With the rise of industry 4.0, the advancement of computers and operating systems require more efficient and optimized scheduling algorithms to assess big data. This study aims to explore and develop a novel and heuristic approach to scheduling algorithms by incorporating predictive models in machine learning, more specifically the linear regression model to predict and allocate the most efficient priority level to each process upon execution. The newly developed ASH-PLR algorithm was tested against common and contemporary scheduling algorithms such as the FCFS, AMRR, and the MMRRA in terms of their Average Turnaround Time, Average Waiting Time, and Context Switches. The results indicate that the ASH-PLR is the superior scheduling algorithm when it comes to processes that have shorter burst times and extensively outperforms the FCFS and MMRRA in terms of Average Turnaround Time and Average Waiting Time. ASH-PLR displays the ability of predictive models to be integrated in future algorithms for better optimizations in upcoming new technology.