Hybrid Gradient Descent and Sea Lion Optimization Algorithm (H-GD-SLnO) to Optimize Task Scheduling in Fog Computing Environment

The architectures associated with cloud and fog computing are fundamentally important for the decentralization of computational tasks, as they facilitate the efficient processing of various user demands while simultaneously supporting extensive Internet of Things (IoT) networks that are increasingly prevalent in today's technological landscape. As the complexity and scale of computational loads continue to expand at a rapid pace, it becomes imperative to implement effective task scheduling mechanisms across these heterogeneous architectures, as this is vital for the optimization of resource utilization and the significant reduction of latency that can adversely affect system performance. This Paper centers on optimization, highlighting the novel integration of Gradient Descent (GD) and the Sea Lion Optimization Algorithm (SLnO). It proposes a harmonious methodology for addressing task scheduling challenges in Fog Computing by utilizing the accuracy of GD for localized searches and the exploratory strengths of SLnO for overarching optimization. The hybrid methodology (H-GD-SLnO) was evaluated using synthetic datasets and real grid workloads, including DAS-2 and AuverGrid, as well as Hadoop tasks like Terasort. The simulation outcomes reveal that (H-GD-SLnO) surpasses numerous pre-existing algorithms, realizing reductions in mean job completion time as well as average scheduling delay and enhancements in resource utilization and throughput. These results emphasize the efficacy of our proposed methodology in meeting the rigorous task scheduling demands within cloud and fog computing environments.