Research On Task Offloading And Resource Allocation Methods In Edge Computing System

With the rapid development of science and communication technology,the data and traffic of the Internet have exploded,and the emergence of resource-intensive and delay-sensitive applications has put forward higher requirements on the computing and storage capabilities of user equipment.In order to alleviate the contradiction between limited user equipment resources and complex applications with high requirements,mobile edge computing emerges as the times require.By deploying servers at the edge of the network,MEC sinks its powerful computing and storage capacity to the edge of the network and provides users with high-performance and low delay network services.However,MEC computing resources are limited and cannot meet the needs of all users.Therefore,how to make reasonable and optimal task offloading decisions and allocate optimal computing and communication resources is of great significance in the research of MEC.This thesis studies the task offloading and resource allocation of multi-users in edge computing systems from two aspects: single-antenna overall offloading and multi-antenna partial offloading.The main studies are as follows:1.Aiming at the overall offloading problem in multi-user single antenna scene,data compression technology is introduced,and a task offloading algorithm based on chaotic binary particle swarm optimization is proposed.In this thesis,the task offloading problem is modeled as a constrained integer programming problem.The CBPSO algorithm is used to solve the 0-1 integer programming problem of the task offloading decision which effectively reduces the delay of task execution.The simulation results show that,compared with all-local offloading,all-MEC server offloading and offloading strategy based on binary particle swarm optimization,the offloading algorithm proposed in this thesis has more obvious advantages in task execution latency.2.Aiming at the problem of partial offloading in multi-user and multi-antenna scene,the data compression technology is introduced,combined with the multiple input multiple output technology,a MIMO-MEC system model is constructed,and a task offloading algorithm based on adaptive particle swarm optimization is proposed.Different from the previous part,the user’s tasks can be partially offloaded,and the communication resources are more comprehensively optimized.In this thesis,the task offloading problem is modeled as a resource optimization problem with constrained delay minimization.Through the proposed offloading algorithm,the reasonable allocation of offloading ratio,compression ratio,transmission power,MEC computing resources and channel bandwidth is realized,which effectively reduces the time delay of task execution.The simulation results show that compared with all-local offloading,all-MEC server offloading and offloading strategy based on PSO,the offloading strategy based on APSO proposed in this thesis has lower latency and effectively improves the system performance.