Simulation of Cloud Computing Resource Allocation Optimization Model Based on Graph Neural Network

Cloud computing technology improves the resource utilization of clusters by integrating and redistributing physical resources. Virtualization technology is the foundation for abstracting, classifying, and integrating physical resources, as well as providing elastic computing services. With the increasing complexity of application scenarios and requirements, the structure of network models is constantly being optimized and updated. In this paper, by offline or online learning of the graph cloud computing resource allocation optimization model, the neural network structure, weights, thresholds, and other parameters are continuously adjusted to achieve the optimal criterion function of a certain error, thereby obtaining the resource allocation optimization model of the modeled nonlinear system. We optimized cloud computing resource allocation based on GNN (Graph Neural Network) and constructed a model. Through simulation experiments, it has been shown that the GNN based cloud computing resource optimization allocation model constructed in this article has more resource allocation than traditional models. The maximum allocation total of the optimization model constructed in this article can be 2800, while the maximum allocation total of the traditional model is 2300. ft can be seen that the optimized model in this article has a greater throughput.