A smart collaborative framework for dynamic multi-task offloading in IIoT-MEC networks

The rapid development of Industrial Internet of Things (IIoT) has brought unprecedented opportunities to the industry informatization. However, facing with billions access of IIoT devices, the traditional IIoT architecture based on cloud computing is no longer suitable in terms of flexibility, efficiency and elasticity. Multi-access Edge Computing (MEC) has been seen as a enabling technology to process massive time-sensitive tasks. Meanwhile, the multi-task collaborative offloading is an urgent problem for IIoT-MEC networks. In this paper, a Smart Collaborative Framework (SCF) scheme is designed to achieve dynamic service prediction and make multi-task offloading decisions. First, a theoretical model, including a Hierarchical Spatial-Temporal Monitoring (HSTM) module and a Fine-grained Resource Scheduling (FRS) module, is established. Hybrid deep learning algorithms are applied to the monitoring module from spatial-temporal dimensions. Besides, both mixed game and improved queuing theories are adopted to enhance offloading efficiency in the FRS module. Second, a specific framework and an implementation process are designed for illustrating scheme details. Third, a prototype environment are created with optimal parameter settings. The validation results demonstrated that the SCF scheme can achieve better task awareness, abnormality inference and task offloading compared to other candidate algorithms. The proposed model has enhanced 7.8% and 8.5% in accuracy and detection rate, and optimized the offloading efficiency.