Research On Computing Offloading And Dynamic Pricing Technology Based On Edge Collaboration

With the gradual popularization of 5g network,mobile edge computing,as one of the main technologies of 5g,has great research value.Mobile edge computing means that local devices can reduce the delay and power consumption of business services by offloading tasks to the server near the network edge for processing.Mobile edge computing mainly includes three aspects: offloading decision,resource allocation and system implementation.Offloading decision is used to determine whether to offload and how much data to offload.Resource allocation is used to determine how MEC server divides resources to process various terminal tasks after offloading.However,due to the influence of hardware and other factors,MEC server resources are limited,so how to maximize the resources required for end users is a hot research topic.In this paper,we consider a MEC server with limited resources.When the terminal has a large demand for data,the first case is to expect the resource demand of the terminal,and the MEC server will offloading some tasks to the nearby MEC server for collaborative processing;the second case is to know the demand of each terminal,and the MEC server selects the appropriate terminal to provide services under the condition of meeting its own limited resources.This paper mainly studies the dynamic pricing and offloading strategy of mobile edge collaborative computing and the allocation of composite resources.The main research work includes:(1)two system models are constructed to meet the requirements of a MEC server with limited resources to provide services for a large number of mobile terminals.The first system model is to unload part of the computing tasks to the cooperative server when the MEC server resources are insufficient to meet the needs of the mobile terminal,so as to maximize the profit;the second system model is to select some terminals to provide services for the MEC server when the MEC server resources are limited,so as to maximize the profit.(2)In the first system model,heuristic algorithm is used to solve the curse of dimensionality in dynamic pricing and unloading.Numerical simulation and analysis show that the proposed heuristic algorithm is similar to the optimization algorithm.(3)The improved genetic algorithm is used to solve the resource allocation problem of MEC server in the second system model,and the numerical simulation shows that the improved genetic algorithm is effective for resource allocation.