Intelligent scheduling task adjustment method based on graph neural network and adaptive weights

In order to make the scheduling of tasks to be executed by intelligent agents more flexible and reasonable, a method for adjusting intelligent scheduling tasks based on graph neural networks and adaptive weights is proposed. Based on the short-term memory ability of graph neural networks, search for temporary computing tasks received by intelligent agents in the database as the basic data for subsequent arrangement and adjustment. Extracting thematic content features of intelligent tasks based on recurrent neural networks; Identify priority task targets in the scheduled tasks through graph neural network models. Based on adaptive weights and combined with online learning listwise algorithm, intelligent task scheduling adjustment is achieved. The experimental results show that the conflict rate of task scheduling adjustment using this method remains stable within 5%, and the scheduling adjustment effect is more reasonable; The utilization rate of computing resources is higher after task scheduling adjustment, and task scheduling is more flexible; The longest response time for task scheduling adjustment is 29ms, which is more effective.