Research On Mobile Edge Computing Offloading And Resource Allocation Strategy

The base station densification and computing capability deployment close to network lead to a trend to combine ultra-dense networks and mobile edge computing(MEC)servers to enhance the system to meet the demands of users for high rate,pervasive connection,low latency,and low power consumption.However,the limited infrastructural resources of MEC and the complex environment of ultra-dense networks result in high performance degradation,which will affect the benefits of computation offloading.Therefore,it is necessary to thoroughly study task offloading and resource allocation problems in ultra-dense mobile edge computing networks.Firstly,this thesis introduces the research background and overview of MEC,describing the collaborative research on task offloading,resource allocation,and content caching.This chapter analyzes the characteristics of existing studies in details,and then summarizes them.Secondly,for deployment scenarios with multiple MEC servers in ultra-dense networks,user terminals and MEC servers are selected for bidirectional matching.The variables of the computing resource allocation and offloading decision are served as controllable parameters to model the weighted to minimize time and energy consumption optimization problem.Furthermore,a sub-optimal solution is leveraged to decompose this NP-hard problem into two sub-optimization problems,namely,computing resource allocation and offloading decision.Based on that,a multi-base station game offloading algorithm is proposed,which uses a Lagrange multiplier method to solve the problem of computing resource allocation,then exploits matching game theory to coordinate the mutual selection between users and servers.And the best match based on the preferences of both parties can be obtained.Numerical results show that our proposed algorithm can effectively reduce system overhead.Thirdly,in MEC-assisted ultra-dense networks with commercial caching services,the users have to obtain the cached content before further processing the task.In order to solve this problem,a joint base station selection,caching decision,offloading decision,and computing resource allocation strategy is developed.Based on that,an ultra-dense architecture based on software-defined networks is designed to realize centralized logic control and simplify network resource management.Furthermore,an intelligent offloading and caching algorithm based on Q-learning is proposed,which can adjust to environmental dynamics through exploration and learning,thereby yielding the best base station selection,caching,offloading,and computing resource allocation strategy.Simulation results show that the proposed algorithm can significantly reduce the average overhead of users.Finally,the research contributions of this thesis are summarized and future research directions are given.