Cloudlet Placement Technology Based On Heuristic Algorithm In Wireless Metropolitan Area Network

With the development of the Internet of Things and 5G technology,mobile applications gradually tend to be computation-intensive,which requires more computing resources for the mobile terminals.At the same time,users’ demand for the timely response capability of mobile devices is also increasing.However,due to the limited size of mobile device,its computing power and battery life cannot meet the needs of users.Mobile edge computing paradigm is an effective way to solve the above problems.Cloudlet is composed of computer cluster with rich computing and storage resources.By placing cloudlet on nearby edges for terminal,edge computing service providers can provide high-quality computing services with lower network delay,so as to improve the user experience.And mobile terminals can utilize the rich computing resource by offloading tasks to cloudlet to make up for the shortage of its own computing resources.However,most of the existing researches focus on computation offloading of mobile applications,ignoring the impact of placement strategies on service quality.So how to optimize cloudlet placement in a given network remains to be further studied.Therefore,this paper proposes two cloudlet placement strategies in WMAN for two types of tasks respectively with the goal of minimizing the system response time,based on two heuristic algorithms.And the effectiveness and superiority of the proposed method are verified through the simulation experiments.This paper has the following contributions:(1)A unified network topology model including mobile terminals and base stations in WMAN environment is established.According to the network topology model,mathematical modeling of the cloudlet placement problem is carried out,and the cloudlet placement model in the edge environment is constructed for different terminal task types.The problem to be solved in this paper is explained from the model level.(2)An offloading scheduling model based on simple task is proposed,which consider the influence of the structure of simple task model on cloudlet placement in edge environment.According to this model,a cloudlet placement decision technology oriented to simple task was proposed.And it can effectively reduce the average response time of the system by improving the traditional particle swarm optimization algorithm.(3)An offloading scheduling model for workflow applications based on complex scientific workflow is proposed,which considering the internal dependency structure of the scientific workflow and the influence of its computing type on the execution efficiency of the cloudlet server in edge environment.According to this model,a discrete particle swarm optimization algorithm based on genetic algorithm operator is proposed.And it can effectively avoid the problem of premature convergence of particle swarm optimization and reduce the average response time of the system by introducing crossover operator and mutation operator of genetic algorithm into the traditional discrete particle swarm optimization.Finally,in order to verify the effectiveness of the proposed strategies,the real-world telecom base station datasets and scientific workflow in different fields were used to simulate the experimental environment.The proposed strategies were evaluated comprehensively according to the experimental results.The evaluation results show that the proposed strategies can effectively reduce the average response time of the system compared with other cloudlet placement benchmark algorithms.