Mobile user deep neural network calculation unloading time delay minimization method

Abstract

The invention discloses a mobile user deep neural network calculation unloading time delay minimization method. A deep neural network-based intelligent application processed by a user in a moving state is analyzed, and a corresponding model is built for minimizing the time delay; and the deep neural network is modeled into a directed acyclic graph, and an optimal unloading decision is made by cutting the graph for multiple times; the cutting process is divided into two stages: in the first stage, only an optimization problem in one time slot is considered, an optimal cutting point is found, anetwork model in the time slot is divided into a front part and a rear part, the first part is subjected to local calculation, and the second part is unloaded to edge cloud calculation; in the secondstage, starting from the global perspective, the whole directed acyclic graph is cut into a plurality of blocks, and the number of time slots required for completing the whole DNN task is minimized. According to the method, the purpose of minimizing the DNN calculation time delay is achieved, cooperative cooperation of the mobile user and the edge cloud server is achieved, and meanwhile the continuity of task processing in the user moving process is guaranteed.

Description

technical field [0001] The invention belongs to the technical field of edge computing offloading, and in particular relates to a method for minimizing offloading time delay of intelligent application computing offloading based on a deep neural network in a mobile state of a user. Background technique [0002] The latest research results of Deep Neural Networks (DNN) have greatly improved the performance of DNN, and have been widely used in image recognition, intelligent search, language processing and other fields. Although the central processors of new mobile devices are becoming more and more powerful, they still cannot meet the delay requirements of DNN intelligent applications with extremely high real-time requirements. [0003] Edge computing is an effective way to solve the above problems. Mobile devices can hand over part or all of the computing tasks to edge cloud servers through computing offloading to solve the shortcomings of mobile devices in terms of resource s...

AI Technical Summary

Problems solved by technology

With the advent of the 5G era, the mobility of users cannot be ignored. Users often change the base stations connected to them during the mobile process, which greatly increases the failure rate of computing offloading.

Images