Research On Computing Resource Allocation And Task Scheduling In MEC System

As a supplement to mobile cloud computing,mobile edge computing sinks cloud computing resources to the edge of wireless access networks to provide closer cloud services for mobile users,reduces the transmission delay of computing tasks,reduces the transmission pressure in back haul networks,and saves The energy consumption of mobile devices.Heterogeneous networks have multiple types of low-power wireless access nodes,which can meet uninterrupted business need.Deploying edge servers on heterogeneous network wireless access nodes are considered an effective way to improve future network performance.In a heterogeneous network environment,multiple wireless access technologies coexist,providing more options on mobile users to computing offloading.At the same time,it also increases the difficulty of allocating bandwidth resources and computing resources in wireless access networks.In this thesis,resource allocation and task scheduling for mobile edge computing system are studied under heterogeneous and ultra-dense network environment.The main works are as follows:1.A computational offload algorithm for joint allocation of bandwidth and computing resources(COA-JBCR)is designed to tackle the problem of users and servers bidirectional selection for resource allocation in heterogeneous network environment.This algorithm takes into account computational offloading,bandwidth resources and computational resource management.Next,this thesis establishes an optimization model on maximizing the total benefit of system tasks execution under the constraint on deadline.this optimization model is transformed into the problem of minimizing the offloading cost and the alliance game method is used to obtain the optimal resource allocation strategy.The simulation results show that the designed algorithm can better improve the system efficiency and reduce the system task execution delay.2.A high response ratio scheduling algorithm based on multi-DAGs(HRSA-MD)is designed to tackle the problem of low resource utilization of distributed heterogeneous edge servers.Next,this thesis establishes a problem optimization model on the goal of minimizing system completion time and proposes a multiple attribute task prioritization scheme proposal.A Dijkstra algorithm is used to select the optimal task transmission path,which makes the total link bandwidth as used to transmit task data as soon as possible.The simulation results show that HRSA-MD outperforms HEFT,CHP and OMC algorithms in reducing system completion time.