Intepretable Deep Gaussian Naive Bayes Algorithm (Idgnba) Based Task Offloading Framework for Edge-Cloud Computing

Abstract

When it comes to Internet of Things (IoT) applications and machine learning based computing, resource-restricted edge devices are inadequate due to the exponential growth of mobile information and the massive need for processing power. An edge offload, the migration of complex tasks from IoT devices to edge cloud servers, is a distributed computing paradigm that has the potential to overcome the IoT device resource limits, lessen the computational load, and increase the effectiveness with which activities are processed. However, due to the NP-hard nature of the optimum offloading decision-making issue, an efficient solution using traditional optimization techniques is difficult. Current deep learning algorithms still have a lot of problems, such as their slow pace of learning and limited ability to adapt to new environments. We provide a unique interpretable deep Gaussian naive Bayes technique (IDGNBA) for extremely fine offloading choices to address these issues. Through several simulation studies, we assess the efficacy of IDGNBA and find that it performs better in terms of offloading than traditional techniques. The model has strong mobility and can quickly adjust to a fresh MEC working atmosphere while taking offloading decisions in real-time.