Multi-node task scheduling algorithm for edge computing based on multi-objective optimization

Abstract

In edge computing, nodes are highly dynamic and resources are unbalanced. Due to the need to ensure near real-time response of tasks, when task scheduling is required, the resource availability of the target node needs to be estimated. It prevents the node's computing resources from being exhausted or the node from failing or going offline, when the task is pushed to the target node. In this paper, by studying multi-node task scheduling technology, a multi-objective optimization model is established, while considering the impact of completion time, energy consumption and load balancing on task scheduling. The task scheduling problem is transformed into a bidding model, and the offloading location of subtasks is determined in real time to meet the requirements of delay-sensitive tasks. Finally, simulation experiments are used to obtain the availability of the technology for multi-node task scheduling, which provides new ideas for task scheduling in edge computing.