A Cloud-Side Task Scheduling Algorithm with Multiple Evaluation Metrics

Abstract

With the popularity of intelligent terminal devices, edge computing has been fully developed. Power patrol robot is widely used in power grid information collection, and edge computing can effectively shorten response time, improve processing efficiency and reduce network pressure, so as to meet the real-time requirements. However, the following problem is how to realize the scheduling strategy of edge cloud and central cloud and optimize multi performance indicators. To solve this problem, this paper proposes a task scheduling model combining genetic algorithm with Docker container technology and taking cloud computing center and edge cloud into comprehensive consideration. Firstly, the task is classified by condition analysis. Assign tasks to cloud computing centers or edge nodes according to the task type; Genetic algorithm is used to assign tasks to edge nodes. Finally, the performance of the model is verified in the simulation environment. The experimental results show that this task allocation method greatly improves the resource utilization of edge server equipment on the basis of considering the needs of tasks, the limited resources of edge server, and meeting the needs of task proposers.