An Online Fog Computing Task Offloading Algorithm Based on Robust Evolutionary Optimization

With the rapid development of the Internet of Things (IoI), a large amount of raw data produced by IoT devices must be processed promptly. However, due to the limited resources of IoT devices, fog computing, and cloud computing platforms are adopted to handle the on-demand requirements through task offloading strategies. Collecting the status of all computing nodes consumes many resources, making it challenging to generate a globally optimal solution online. In this work, we propose a scheme that combines offline uncertain optimization and online scheduling to produce a high-performance online schedule based on the offline uncertain 'IoT-Fog-Cloud' task scheduling solutions. We define the offline uncertain task scheduling and online scheduling sub-problems and design an online strategy to optimize the overall utilities of tasks based on the existing fuzzy scheduler. The simulation results demonstrate that our algorithm performs better on task utilities and its robustness index compared to the heuristic method.