Adaptive Task Scheduling in Fog Computing Using Learning Automata and RBF Neural Networks for Optimized Performance and Energy Efficiency

Fog Computing (FC) acts as an intermediate computational layer between the cloud and Internet of Things (IoT) devices, designed to enhance service quality by processing tasks closer to the data source. However, effectively managing energy consumption (EC) remains a critical challenge due to the complexities of task scheduling. This paper proposes an enhanced task scheduling approach based on learning automata (LA) and neural network modeling to minimize fitness, makespan (MK), and associated costs in fog environments. Furthermore, an additional radial basis function (RBF) model is introduced to predict interdependencies among MK, fitness, and cost relative to virtual machine (VM) configurations. A Comparative analysis demonstrates the superior performance of the proposed LA-driven scheduling model over existing methods, achieving more efficient resource allocation and environmental impact reduction across key metrics. This study advances FC task scheduling techniques, highlighting the potential of integrated neural network models to optimize energy-aware computation.