Research On Optimal Scheduling Of Edge Computing Resources For Distribution Internet Of Things

In recently years,5G technology is developing fast,the current distribution network is developing in the direction of information and intelligent distribution Internet of things(D-IoT).The distribution Internet of things adopts the mode of collaborative work between cloud computing center and edge cloud network,in which the edge cloud network has a large number of intelligent terminals,which can well adapt to the current rapid expansion of power data scale.How to allocate the limited computing resources to make the distribution Internet of things run more efficiently is a hot issue in mobile edge computing(MEC).In order to further improve the ability of edge network in D-IoT to process data in real time,this dissertation proposes an edge computing task unloading strategy based on dynamic non cooperative game,which optimizes the unloading of computing resources from the cloud computing center to the intelligent terminal of the edge network,and then formulates the strategy of power distribution IOT The computing resource allocation strategy based on ea-moalo algorithm for the edge server in the interconnection system optimizes the allocation of computing resources from the intelligent terminal to the adjacent edge server.and finally realizes the optimal scheduling of the edge computing resources in the distribution Internet of things.The special contribution is as follows1.We studied the network framework of D-IoT.Aiming at the four layers’ structure of D-IoT,a novel optimal scheduling scheme of computing resources is proposed.The scheme can effectively adapt to the characteristics of the distribution Internet of things,such as the numerous computing tasks and the dynamic changes at all times,we reduced the burden of the cloud computing center successfully,and improve the system efficiency of the D-IoT.2.Research on the strategy of edge computing task unloading based on dynamic non cooperative game.According to the characteristics of D-IoT,this dissertation proposes a real-time optimization goal which considers the system energy consumption and user experience(QoE),and regards the behavior of edge network requesting computing resources as a non cooperative dynamic game behavior.In this dissertation,the calculation process is decomposed into a series of distributed sub optimization problems.Aiming at solving the potential problem,we proposed a distributed optimization algorithm.Finally,a series of experiments show that the algorithm we proposed in this dissertation is better than the traditional centralized optimization method in solving the calculation unloading problem of D-IoT.3.The resource allocation strategy based on ea-moalo algorithm is studied.The traditional moalo algorithm has some defects,so it can’t fully apply the resource allocation strategy of MEC.Therefore,this dissertation combines the idea of differential evolution and adaptive competition mechanism to improve the original moalo algorithm,so as to avoid the algorithm falling into local optimum and accelerate the convergence speed.Considering the access delay,throughput and energy consumption,we established the system optimization model,and the total system welfare is defined as the index to measure the resource allocation efficiency of IOT system.The results of numerical analysis show that the proposed resource allocation strategy based on ea-moalo algorithm is effective.