Task and data management optimization method based on graph neural network

The invention relates to the technical field of cloud computing, in particular to a task and data management optimization method based on a graph neural network, and the method comprises the steps: fusing node information of different weights based on a hypergraph, obtaining output node features through the aggregation of related hyperedge features, and obtaining a node classification result; and obtaining a data center where the task is executed. And then obtaining a transmission path and a transmission sequence of data files required by the data task according to a principle that the transmission time is shortest and the number of required files is maximum. Compared with a traditional centralized single data center task processing method, the method has the advantages that complex task requirements submitted by a user can be met, waiting time during task processing is shortened, and the workload of the data center can be reduced; compared with a common distributed data center, under different task numbers a