Bidirectional Gated Cycle Units Enhanced Convolutional Network for Cloud Energy Optimization

Abstract

The rapid expansion of cloud computing has led to unprecedented growth in energy consumption across data centers, necessitating innovative methods for energy optimization. This paper introduces a novel deep learning framework, Bidirectional Gated Cycle Units Enhanced Convolutional Network (BGC-CN), designed to enhance energy prediction and optimization within cloud infrastructures. By integrating bidirectional gated cycle units with convolutional neural architectures, the proposed model effectively captures temporal dependencies and spatial correlations in large-scale cloud workloads. Experimental evaluations on benchmark cloud datasets demonstrate significant improvements in prediction accuracy and resource utilization compared to traditional recurrent and convolutional approaches. The results reveal that BGC-CN achieves a reduction of up to 18% in energy wastage while maintaining service level agreements (SLAs), highlighting its potential to contribute to sustainable cloud computing environments.