In this paper, we consider a computation offloading problem in a fog-enabled Internet of Things (IoT) network consisting of helper fog nodes and task nodes with energy harvesting capabilities. Considering the unpredictable energy supply and the low conversion efficiency in the case of energy harvesting, offloading scheduling plays a critical role to maintain the quality of experience (QoE) for computation intensive applications. Our goal is to find a scheduling policy to minimize the task abandoning ratio subject to the constraints on the energy causality and the queue stability. We propose an asymptotically optimal algorithm to solve this stochastic programming problem. In particular, the task admission, the local execution, and the offloading decisions are jointly optimized. Theoretical analyses are further presented to establish the performance guarantees of our proposed algorithm. Meanwhile, numerical simulations corroborate the optimality of the proposed policy.
Task Offloading Policy for Nodes with Energy Harvesting Capabilities
2019-09-01
209722 byte
Conference paper
Electronic Resource
English