Dynamic Willingness-Driven Collaborative Task Offloading and Resource Optimization in D2D-Assisted MEC for IIoT

Mobile Edge Computing, as an extension and complement to cloud computing, has been introduced into Industrial Internet of Things (IIoT) to significantly reduce decisionmaking latency, save bandwidth resources, and enhance privacy protection. Nevertheless, the resource constraints of IIoT devices and the limited capabilities of MEC servers still pose challenges to continuous local execution. Therefore, constructing network models suitable for IIoT scenarios and optimizing the allocation of shared resources have become pressing issues to address. Considering the resource heterogeneity of IIoT devices where resource-constrained and resource-idle devices coexist, we develop a Device-to-Device (D2D)-assisted MEC offloading model that incorporates the willingness of idle devices to participate in offloading. Our objective is to minimize the total system cost through optimized task offloading and resource allocation. Therefore, We develop a D2D-assisted MEC offloading model with dynamic willingness variations and propose a hybrid approach named HAJTORA that combines the improved ALSHADE algorithm with convex optimization methods for joint task offloading and resource allocation problem. Simulation experiments demonstrate that our algorithm outperforms benchmark methods, including fast convergence and low system cost.