Research On Edge Resource Scheduling In Heterogeneous Wireless Networks

Mobile Internet has played an important role in facilitating people’s lives and promoting industrial development.It is a key development direction of China’s "Internet Power" and "Internet Plus" strategies.With the expansion of network scale and the development of various communication technologies,mobile Internet presents the trend of surging traffic and heterogeneous network development.How to support large-scale,high bandwidth and low delay network service under the environment of heterogeneous wireless network has become the core problem facing the development of mobile Internet.In view of the above trend,an effective solution is to efficiently utilize the edge resources of the network.As is known to all,edge nodes are closer to users.If network services can be provided by edge nodes,it can not only reduce network service delay and improve network service quality,but also reduce data traffic transmission in the network and network load pressure.Therefore,many scholars have carried out researches on the optimization of edge resources such as cache space,multicast resource,computing capacity.However,current studies still face a series of problems and challenges,which are as follows:(1)The existing content caching passively stores the service content after user request,resulting in the caching decision lagging behind the user request.(2)The existing multicast service scheduling methods neglect the constraint of storage space and the collaboration between multiple nodes,and do not carry out differentiated scheduling according to the difference of service requirements.(3)The existing computing offloading methods usually ignore the invisible limitation of the storage space,and do not accurately give the offloading status of each service.(4)The existing computing resource allocation methods ignore the correlation between multicast technology and edge computing,which leads to the lack of research on the allocation of computing resources under multicast scenariosIn view of the above problems and challenges,this paper focuses on heterogeneous wireless networks which consists of small base stations,macro stations and remote servers,and studies the problems of content caching,multicast service scheduling,computing offloading,computing resource allocation to reduce the network delay of service and improve the quality of the user experience,specifically including:(1)In the term of edge cache space optimization,a knowledge-based active caching strategy according to user demand cognition is proposed.We first analyse user request record history through the deep belief network,and set up the forecast model of user request.On this basis,an active caching algorithm is proposed based on greedy algorithm,which efficiently improves the cache hit rate and reduces the service delay.(2)In the term of multicast service scheduling,a green and efficient random cooperative multicast scheduling strategy is proposed.We first build the service demand priority queue,and propose two online multicast service scheduling methods to cooperate multiple base stations.Then,we prove the superiority of proposed algorithms by theoretical proof and experimental verification.(3)In the term of computing offloading,a method based on multi-update reinforcement learning is proposed.Considering the constraints of computing resources and caching resources,the computing offloading problem is constructed as a Markov decision process.We propose an innovative multi-update reinforcement learning method to greatly reduce the complexity of the algorithm and improve the training speed and accuracy of the model.(4)In the term of computing resource allocation,a multicast aware joint optimization method for caching and computing is proposed.We set up a joint optimization model to optimize the service delay,and transform the model by stochastic optimization method.Further,we design an iterative descent method to solve the problem.Finally,the advantages of the algorithm in network energy consumption,service delay and other performance are verified by simulation experiments.This paper forcuses on the edge resource optimization of heterogeneous wireless network,respectively from the caching space,multicast resources,computing capacity and other aspects of the study.We carry on the thorough discussion on content caching,multicast scheduling,computing offloading and multiple resources optimization problem.In terms of research methods,this paper adopts the ideas of problem analysis,mathematical modeling,algorithm design,theoretical proof and experimental verification to improve the scientific nature and reliability of the research.The results of this paper have certain reference significance for promoting the research and development of edge resource optimization in heterogeneous wireless networks.