Mobility- and Energy-Aware Cooperative Edge Offloading for Dependent Computation Tasks

Cooperative edge offloading to nearby end devices via Device-to-Device (D2D) links in networks with sliced computing resources has mainly been studied for (helper nodes) that are stationary (or follow predetermined mobility paths) and independent computation tasks. However, often mobile, a given application request commonly requires set of dependent We formulate novel model the cooperative tasks mobile helper nodes. task dependencies general dependency graph. Our employs state-of-the-art deep-learning-based PECNet offloads only when sojourn time coverage area node or Multi-access Edge Computing (MEC) server is sufficiently long. minimization problem consumed battery energy execution, data transmission, waiting offloaded results on devices. convert resulting non-convex mixed integer nonlinear programming into an equivalent quadratically constrained quadratic (QCQP) problem, which we solve Energy-Efficient Task Offloading (EETO) algorithm. The numerical evaluations indicate EETO approach consistently reduces consumption across wide range complexities completion deadlines can thus extend lifetimes operating resources.