A Smart Mathematical Approach to Resource Management in Cloud Based on Multi-objective Optimization and Deep Learning

Scaling resources in Cloud computing has become complex and challenging as it has unlimited resources and works on “pay as u go” mode to meet user demands. To ensure that resources are allocated efficiently, Cloud services can be highly available, but they are still susceptible to failures due to the complex and dynamic nature of virtual machine allocation in the cloud environments, in other words, it is difficult to manage and control resources or to choose the best allocation of these resources. Our purpose in this paper is to solve one of the major issues faced by cloud computing, namely resources management in cloud environments. To deal with this issue we need a significant aspect of task scheduling in cloud, such that load balancing as it offers a huge aid to perform task management in the cloud. Therefore, we propose a smart mathematical approach that aims to rebalance the cloud environment by modeling it with a matrix. It is based on arithmetic constraint to balance the load of virtual machines between host machines in cloud environments. The proposed algorithms seek to balance the cloud system matrix by following an emigration trick of virtual machines, we ended the approach by proposing new multi-objective model to follow in order to obtain the rebalanced matrix modeling the cloud environment after the rebalancing process.