Conditional Adversarial Domain Adaptation Framework with African Buffalo Optimization for Robust Load Balancing in Cloud Computing Environment

Cloud computing systems often encounter challenges in load balancing and performance optimization due to dynamic workloads and heterogeneous resources. This study proposes a hybrid framework integrating Conditional Adversarial Domain Adaptation (CADA) with African Buffalo Optimization (ABO) to enhance task scheduling, load distribution, and energy efficiency in cloud environments. The CADA framework enables effective knowledge transfer between source and target domains by aligning feature distributions through adversarial learning, while ABO optimizes task allocation by minimizing energy consumption and latency and maximizing throughput. Simulations using synthetic cloud task datasets with 100 nodes demonstrated significant improvements over baseline models. The proposed CADA–ABO achieved a 35% increase in data packet delivery, 28.4 J of residual energy (compared to 12.6 J for LEACH), and a 25% reduction in end-to-end delay. Additionally, network lifetime increased by 30%, and throughput improved by 22%, with domain adaptation accuracy exceeding 88%, confirming strong generalization capability. Overall, the CADA–ABO framework ensures robust, energy-efficient, and scalable load balancing, delivering reliable performance optimization for modern cloud computing systems.