A Novel Machine Learning-Enhanced Swarm Intelligence Algorithm for Cost-Effective Cloud Load Balancing

Cloud computing refers to the instant accessibility of shared computer resources over the internet on-demand, thereby revolutionizing the way organizations deploy and manage their IT infrastructure. Efficient load balancing in a cloud environment scales up, utilizing resources with optimized cost-effectiveness. This paper introduces a new Machine Learning-guided Improved Swarm Intelligence Algorithm for reliable load balancing in cloud computing. Methodology It spans on three main elements. The first is the data center controller for the management of tasks. VM allocation uses an Improved Beetle Swarm Optimization Algorithm and one-dimensional convolutional probabilistic neural network is used for host load prediction. To propose an approach that looks to optimize the key metrics such as make span, average cost of VM, resource utilization, and value response time through iterative optimization process. The results of the experiments reflect the success of the proposed algorithm, because an improved performance of load balancing is obtained in comparison with those achieved using PSO, Grey Wolf Optimization, and Genetic Algorithm. The convergence diagram displays the iterative optimization process achieved by the proposed algorithm in obtaining a balance among various performance metrics. Overall, the methodology proposed in this work will provide promising facilitation in the increased efficiency, scalability, and cost-effectiveness of environments based on cloud computing with a reliable load balancing.