Performance Optimization Of Mobile Wireless Access Network Based On Edge Computing

Recently,smart mobile devices develop rapidly.Computation-intensive applications appear such as augmented reality(AR),face recognition and so on.These applications require much computation capacity and memory of mobile devices,but mobile devices are resource-limited and unable to execute complex applications or several applications simultaneously.Mobile edge computing solves this problem by offloading computation tasks of mobile devices to nearby base stations,wireless access points or cloudlets,which execute computation tasks and return results to users.Mobile edge computing reduces response delay of users,overloads of core network,and improves the utility of edge devices efficiently.This thesis investigates the performance optimization of mobile wireless access network based on edge computing.The main research and innovation include:firstly,a software-defined computation offloading cooperative control framework in wireless access networks is proposed which makes computation offloading decisions according to different optimization objectives and achieves the flexibility and scalability of control.Secondly,delay optimization of computation offloading in multi-hop ad hoc networks is investigated as a use case based on the proposed framework.The delay optimization problem is formulated as a general integer linear programming and a delay-optimization based ant colony optimization algorithm(DoACO)is proposed to solve the problem.Thirdly,as the second use case of the framework,computation offloading oriented optimization of mobility management in heterogeneous wireless access network is investigated.An optimization of average data delivery cost and a joint optimization of data delivery cost and load balance is formulated,and two algorithms including DDCOA and JoDLOA are proposed to solve those two problems.Simulations show that the algorithms proposed perform better than other baseline algorithms.