A NOVEL HYBRID ALGORITHM COMBINING NEURAL NETWORKS AND GENETIC PROGRAMMING FOR CLOUD RESOURCE MANAGEMENT

Abstract

Maximizing performance and cost-effectiveness in cloud computing systems depends on good cloud resource management. To enhance the administration of cloud resources, this work presents a new hybrid method combining Tree-based Genetic Programming (GP) with Convolutional Neural Networks (CNNs). The CNN component shines in feature extraction from complicated, high-dimensional data including system measurements and previous usage patterns. The Tree-based GP then uses these characteristics to evolve complex resource management policies and strategies by symbolic regression. The hybrid model uses the GP's strength in producing adaptive, interpretable rules and the CNN's capacity to recognize complex patterns. Dynamic workload fluctuations, scalability, and cost optimization are among the major issues in cloud resource management that our method tackles. Experimental results show that the suggested hybrid Strategy beats conventional techniques regarding resource allocation efficiency and general system performance. This work opens the path for more affordable and effective cloud computing systems by offering a solid framework for creating intelligent and flexible cloud resource management solutions.