A Hybrid Population-Based Method for Scheduling Multiprocessor Tasks on Two Dedicated Processors
Fatma Zohra Baatout, Naby Doumbouya, Mhand Hifi · 2024
This paper discusses a novel hybrid population-based method for scheduling multiprocessor tasks on two dedicated processors. Combining a modified grey wolf optimizer with key enhancements, it yields a robust approximate algorithm. Initialization uses a carefully selected combination of a greedy sequence and configurations, ensuring solution feasibility through a tailored sigmoid function. A straightforward local operator intensifies the search space, while a drop and rebuild operator counters premature convergence. Benchmark evaluations and comparative analysis discuss its effectiveness, highlighting the method’s ability to discover new bounds compared to recent state-of-the-art approaches.