Edge server joint task unloading and convolutional neural network layer scheduling method

Abstract

The invention discloses an edge server joint task unloading and convolutional neural network layer scheduling method and belongs to the technical field of wireless communication. The method comprisesthe following steps of S1, modeling user equipment task variables; S2, modeling edge server variables; S3, modeling multilayer CNN (Convolutional Neural Network) models; S4, modeling task total delay;S5, modeling a task unloading and CNN layer scheduling constraint condition; and S6, determining an edge server task unloading and CNN layer scheduling policy, and minimizing the task total delay. According to the method, a delay demand of user equipment when the user equipment performs tasks and the task processing performance of the edge servers are taken into overall consideration, the multilayer CNNs deployed in the cloud server are deployed to the edge servers closer to the user equipment, and on the basis of the service capability of the edge servers, partial layers of the multilayer CNNs are called to preprocess the tasks of the user equipment, so the quality of service of users is improved, and the network performance is optimized.

Description

technical field [0001] The invention belongs to the technical field of wireless communication, and relates to an edge server joint task offloading and convolutional neural network layer scheduling method. Background technique [0002] With the rapid development of the mobile Internet and the popularization of smart terminals, applications such as augmented reality (Augment Reality, AR), virtual reality (Virtual Reality, VR) and mobile high-definition video have an increasing demand for Quality of Service (QoS). high. However, the insufficient processing power of smart user equipment and the performance limitations of traditional Mobile Cloud Computing (MCC) technology make it difficult for the network to meet the business needs of users to process large amounts of data in a short period of time. In response to this problem, mobile edge computing technology emerged as the times require. By deploying an edge server at a base station close to the smart user device, and using t...

AI Technical Summary

Problems solved by technology

[0004] At present, studies have considered the deployment of multi-layer CNNs on edge servers. Some literatures have proposed a CNN layer scheduling scheme based on maximizing the load on edge servers. However, the existing schemes rarely consider the problem of delay optimization, which leads to severe limitations in network transmission performance. In addition, , less research considers the association strategy between edge servers and user tasks, making it difficult for the proposed algorithm to optimize network performance

Images