Research On Task Offloading Strategy Based On Mobile Edge Computing In Ultra Dense Network

With the rapid development of mobile Internet and Internet of Things,emerging computing intensive applications such as augmented reality,autopilot and real-time interactive games.These applications put forward higher requirements for delay and energy consumption of mobile devices.Limited by the computing power and battery capacity of mobile devices,the existing mobile devices can not meet the needs of the above applications.Mobile edge computing(MEC)provides a new solution.Mobile edge computing sinks computing resources and storage resources to the edge of the network,and allows mobile devices to offloading the computing intensive tasks to MEC servers,so as to effectively reduce task completion delay and energy consumption.At the same time,as the key technology of 5G,the ultra-dense network can provide users with huge access capacity through intensive deployment of base stations.However,there are some problems,such as the complex and heterogeneous network environment of ultra-dense network,the limited computing resources of MEC and the scarcity of spectrum resources.Therefore,considering the characteristics of ultra dense network and MEC,in the scenario of coexistence of MEC and ultra dense network,formulating reasonable task offloading and resource allocation strategy has become the key to improve system performance.This paper mainly studies the resource allocation and offloading strategies under the UDN scenarios.The main contributions of this paper are as follows:(1)Aiming at the scenario of a multi-user and singe-server in mobile edge computing,a scheme for jointly optimizing computation offloading decision and MEC server resources is proposed.Through the joint optimization of user offloading decision and resources allocation in ultra-dense network,the total task completion delay is minimized.In the process of solving the objective function,the variable is replaced according to the variable relationship to simplify the problem,and then the simplified problem is decomposed into subproblems.Firstly,it is proved that the MEC computing resource subproblem is a convex optimization problem and solved by KKT condition.Then,the channel allocation algorithm based on improved differential evolution algorithm is used to solve the NP hard problem of channel resource allocation.Compared with the benchmark scheme,the proposed algorithm can effectively reduce the total task completion delay.(2)Considering the limitation of single server,on the basis of the first point,it is further extended to the ultra-dense network scenario of multi-user and multi MEC.On the premise of ensuring the maximum tolerable delay of tasks,a joint optimization problem aiming at minimizing user computing overhead is formulated to jointly optimize task offloading,MEC computing resources and user transmit power.Specifically,in order to simplify the solution of the objective function,the original problem is decoupled into two subproblems: offloading strategy and resource allocation.Firstly,the computing resources are allocated proportionally according to the task attributes,and an offloading decision algorithm based on DQN is designed to optimize the offloading strategy.Then,the resource allocation problem is decomposed into MEC computing resource allocation subproblem and user transmit power allocation subproblem,It is solved by convex optimization theory,KKT condition and golden section method.Simulation results show that the proposed algorithm can effectively reduce the total cost of user computing,and its performance is significantly better than other algorithms.