Research on GRU-Based Load Prediction Method for Vehicular Network Servers in a Multi-Task Offloading Environment

In a multi-task offloading environment, to improve the efficiency of resource allocation and utilization in vehicular network servers, this paper proposes a load prediction method based on GRU-RNN. Initially, a GRU-RNN neural network is selected to build the load prediction model, using the Adam algorithm as the optimizer. Subsequently, the Alibaba Cluster-trace-v2018 dataset is utilized, with data normalization and preprocessing, followed by model hyperparameter acquisition through grid search and cross-validation methods. Finally, simulation results show that the trained model can accurately predict the vehicular network server load, demonstrating high accuracy, stability, and reliability.