On The Study Of Resource Management For Mobile Edge Network

By deploying the cache and computation resource at the wireless network edge(e.g.base stations and users)to reduce the distance between users and contents or services,the mobile edge network can reduce the end-to-end delay and backhaul traffic,which becomes the one of main technical topics of the fifth generation mobile communication systems.The mobile edge network has two typical scenarios:edge caching network and edge computing network.They respectively deploy the cache and computation resource at the network edge to achieve the efficient content transmission and satisfy the computation demand.However,due to the mutual effect between communication and cache(or computation),as well as the differentiated capability and load among different edge nodes,the resource management in the mobile edge network will be more challenging.How to achieve the multi-type resource coordination and node cooperation is a major problem in the mobile edge network.This dissertation will investigate the resource management for the mobile edge network,focusing on the user association and cooperative caching of the edge caching network,as well as the Quality of Service(QoS)guarantee and load balance of the edge computing network.Detailed contributions are summarized as follows:(1)Thejoint content placement and user association in the edge caching network:In the edge caching network,taking into account the wireless channel quality,the communication load and the backhaul condition,the joint caching and user association problem is formulated to minimize the average download delay.The problem is proven to be NP-Hard based on a reduction to the fa-cility location problem,which is solved with the aid of McCormick envelopes and Lagrange partial relaxation method.Simulation results show the proposed scheme can reduce 22%of the average delay,compared to the MPC-MS(Most popular content-Maximum SINR)scheme.(2)The performance analysis and cooperative caching placement in the cache-enabled D2D underlaid cellular network:In a cache-enabled D2D underlaid cellular network,the achievable average ergodic rate is firstly derived with the consideration of the contact duration based on stochastic geometry.Then,the cooperation of BS caching and D2D caching is considered.In this complex network scenario,the closed-form expression for successful transmis-sion probability is derived using stochastic geometry.On this basis,a coop-erative caching placement problem is formulated to maximize the successful transmission probability,which can be transformed equivalently into a bicon-vex problem.Then an algorithm is proposed based on block coordinate descent.Simulation results demonstrate the performance gain of the proposed cooper-ative caching policy.Furthermore,simulation results also show that BSs and users do not necessarily need to cache the most popular contents and their co-operation can achieve the higher gain.(3)The performance analysis and resource allocation in the edge com-puting network with the QoS guarantee:In the mobile edge computing(MEC)network,the task offloading process is firstly modelled into a two-stage tan-dem queue model consisting of the transmission queue and computation queue.Based on the model,the effective capacity of MEC network under the QoS guar-antee is derived to measure the task processing capability.The analytic results show that the bandwidth and computation resource exhibit the bottleneck effect in terms of the supported arrival rate,and exhibit complementation with each other in terms of the end-to-end delay.Then the joint optimization problem of resource allocation and user admission is formulated with the QoS constraint.T his problem is proven to be NP-hard by the reduction to the two-dimensional knapsack problem,which is solved by using the alternating direction method of multipliers.Simulation results demonstrate the proposed scheme can achieve better performance than other schemes by adjusting the communication and computation resource and choosing the optimal subset of accessed users.(4)The task offloading and resource allocation in the three-tier com-puting network with edge node cooperation:In a three-tier computing net-work consisting of devices,edge nodes and cloud servers,leveraging the coop-eration among edge nodes achieves the computation load balance.In this net-work,subject to the limited battery capacity of devices,the offloading decision and the computation resource allocation is jointly optimized to minimize the av-erage task duration.The problem is proven to be NP-Hard based on a reduction to the 3-SAT problem.To solve the problem efficiently,a parallel optimization framework is proposed by utilizing reformulation linearization technology,al-ternating direction method of multipliers and difference of convex functions programming.Simulation results demonstrate the proposed scheme can obtain a near-optimal performance with the low complexity,and can dramatically re-duce the running time.The running time increases slow ly with the increasing number of base stations and users.Furthermore,compared with the scheme without the consideration of the edge node cooperation,the proposed scheme can reduce up to 24%of the task duration.