Research And Design On Resource Scheduling For Weak Computing Capabilities At The Edge

With the advent of the 5G era,massive amounts of Io T devices are connected to the edge of the network,and the tasks generated by these devices require faster data response and more adequate computing resources.To meet these demands,resource scheduling techniques are widely used in edge and cloud computing environments to improve resource utilization and performance performance.Resource scheduling techniques can dynamically allocate and manage compute and storage resources based on user demand and resource availability to achieve the fastest response time and highest performance performance.In edge computing and cloud computing environments,resource scheduling algorithms can allocate resources based on task types and priorities,while dynamically adjusting resource allocation based on resource usage and load to improve the efficiency and stability of the system.To achieve optimal resource scheduling and efficient utilization of the computing resources at the edge devices,this paper presents the following main contributions:(1)A joint optimization problem for task scheduling,offloading decision,and resource allocation in a multi-MD(Mobile Device)and multi-MEC(Mobile Edge Computing)server integrated system is studied,where a support deep Q-learning task deployment and resource allocation algorithm is proposed for task offloading in MEC systems.The algorithm efficiently allocates resources and schedules tasks under CPU frequency,transmission power,computing resources,and time delay constraints,thereby achieving the goal of minimizing the weighted sum of server workload and energy consumption.(2)In a cloud-edge-end architecture composed of multiple MDs,multiple MEC servers,and multiple cloud servers,an adaptive resource allocation algorithm based on task priority is proposed to address the resource allocation problem for different priority tasks.Firstly,local preprocessing is performed on tasks,and tasks are divided into delay-sensitive tasks and delay-tolerant tasks based on their maximum waiting delay and task type.Secondly,different resource allocation schemes for different task types are discussed under different available computing resources of edge servers,and the optimal solution is obtained through mathematical modeling.The experimental results show that the proposed algorithm not only achieves convergence within a small number of iterations,but also effectively improves the resource utilization of the edge server,reduces the waiting time of high-priority tasks,and achieves significant results in balancing energy consumption and server load and shortening the average execution time of tasks,which can better meet the resource allocation and task deployment requirements of large-scale mobile edge computing systems.It can better meet the requirements of resource allocation and task deployment for large-scale mobile edge computing systems.