Multi-task offloading in mobile edge computing networks

Abstract

In order to achieve the optimal balance between task execution delay and energy consumption in Mobile Edge Computing (MEC) networks. First, the Analytic Hierarchy Process (AHP) is adopted to classify the priority of all tasks, so as to establish a related model for task offloading strategy and weight allocation of resources. Then, a multi-task offloading algorithm based on DNN is introduced to generate offloading strategies using multiple DNNS. Meanwhile, training samples composed of offloading strategies and input data are stored through the experience pool. These training samples will be used to train DNN. Simulation results show that the accuracy of the proposed multi-task offloading algorithm can reach 0.99, and the total delay of task processing and system cost can be effectively reduced compared with the three comparison algorithms.