Edge computing network task unloading method equipped with cache auxiliary device

Abstract

The invention discloses an edge computing network task unloading method equipped with a cache auxiliary device. The method comprises steps of establishing execution mode of a user task into differentunloading decision variables to compute energy consumption of the user task in different execution modes such as local CPU processing, edge server execution and content cache processing; constructingan energy consumption value function for minimizing the worst case of a user during task execution, converting the energy consumption value function into a standard QCQP model, and obtaining a cache-assisted pre-selected unloading set by using a semi-definite relaxation and random probability mapping method; and finally, optimizing the transmission power and edge computing resource allocation in the preselected unloading set by adopting a Lagrange dual decomposition method and a dichotomy, and further obtaining an ideal unloading decision by comparing the equipment energy consumption in the set.

Description

technical field [0001] The present invention relates to a joint unloading method of multiple binary variables, in particular to a method for task unloading and resource optimization of user data when adding content caching decisions in a mobile edge computing network system. Background technique [0002] Computing offload technology is a new and efficient data task processing method developed in 5G Mobile Edge Computing (MEC) network, and it is an important means for MEC system to realize real-time execution of terminal applications. As the expansion and extension of cloud computing, edge computing can distribute the computing tasks of the cloud. Therefore, computing offloading can offload all or part of the computing functions to the MEC server near the user end. Its main operation process includes offloading decision-making and offloading information execution. , The data result is sent back to 3 parts. Among them, the offloading decision is made by the MEC control center...

AI Technical Summary

Problems solved by technology

The problem with this invention is that the task feature perception module does not submit the energy consumed by the user's offloading tasks, so when the information analysis and processing module evaluates the performance of the best candidate offloading MEC server, the size and type of the data task will affect this Uninstall method has no small impact

It should be pointed out that patent CN110096318A and patent CN110087257A are based on the discussion of fixed edge server capabilities, and do not consider decision-making under other auxiliary computing conditions such as content caching, and the actual application range is narrow and the scalability is low.

[0004] In the patents based on edge computing task offloading, the traditional method does not consider more offloading factors, the actual application range is narrow, and the scalability is low

Images