MITIGATING STARVATION AND BALANCING FAIRNESS IN REAL-TIME SYSTEMS WITH A HYBRID MULTI-PRIORITY FCFS SCHEDULER

Abstract

CPU scheduling is crucial in determining how a system prioritizes and manages workloads. Effective CPU resource allocation optimizes performance, maintains fairness, and prevents prolonged delays or starvation of lower-priority tasks. Poor strategies can cause bottlenecks and increased response times, making algorithm choice a critical factor in system design. This study proposes a novel approach built on First-Come-First-Serve with an added multi-priority mechanism. A key aim for this approach is the ability to dynamically adjust priority, mitigating starvation and improving fairness. Unlike many priority scheduling methods that rely on static priority assignments, this algorithm continuously adapts to queue conditions and elapsed wait times; ensuring no process remains indefinitely delayed due to priority constraints. The results demonstrate a consistent improvement to response times compared to First-Come-First-Serve, reducing response times by 50.59%. As a result, this method is well-suited for real-time systems and environments requiring quick task execution. It improves real-time system performance while maintaining the algorithm’s inherent short-burst efficiency and straightforward implementation.