A Multi-Objective Optimization Scheduling Method Based on the Genetic Algorithm in Cloud Computing

Introduction

Cloud computing has taken on a certain measure of invulnerability and provides excellent job security, because of its decentralization and technological characteristics. As cloud computing is put together with the following characteristics, it provides an advanced network environment, scale, fast responsiveness, virtualization, geographical delivery and mobility, and new web-based learning capabilities. This gives robust computing. Usage of cloud computing permits users to provision, process, store, and network essential software tools, virtualization software, web-based platforms, apps and databases, web servers, operating systems and application platforms to be accessed and used from the cloud. Cloud computing is becoming more widespread, it's a great way to store and use data. Advanced protection and distributed scope are other cloud computing features and additional benefits. Users can use cloud computing to access, operate, manage, store, and store platforms for all of the various computing tools in the cloud. The cloud computing environment yields several benefits: cost reduction, scalability, energy efficiency, and rapid adoption. The services used by the cloud providers are referred to as software as a service (SaaS), platform as a service (PaaS) and infrastructure as a service (IaaS). These services are delivered to users of dynamic and scalable cloud infrastructure, which are ultimately open to unrestricted use. A user's provisioned virtual resources enable cloud computing resources to be more versatile and elastic. IT services are paid on a pay-as-you-use basis, but user flexibility has widespread cloud use. The major component of this study will be the use of IaaS cloud technology, where computing resources are provided as a service.

To expand resources, tasks should be set on VMs at optimum intervals to maximize the use of physical resources. As with cloud computing environments, the key goal is to get the profit from the computing resources. Various scheduling algorithms are involved in the optimization process. To achieve this requirement, users' needs must be considered while employing an effective scheduling algorithm. One key aim of scheduling tasks is balancing the processing load across virtual machines (VMs) while also maximizing resource usage. In task scheduling, the two main aims are to distribute the processing load among VMs while at the same time maximizing their resource utilization and to minimize the total execution time. User and resource needs both affect task scheduling. There is a trade-off in that one has to prioritize the other. Though tasks and resources vary, some tools are consistent for all users, allowing more complex configuration management capabilities. Since tasks should be given enough time to complete, it is desirable to develop an allocation algorithm that improves the possibility that each server has sufficient resources to accomplish their tasks. The main goal of a scheduling objective is to manage tasks to be completed under the constraints that they serve to meet a particular objective.

This paper proposed a multi-objective improved genetic algorithm (MOIGA) to address multi-objective task scheduling problems. The MOIGA algorithm uses a multi-criteria objectives made up of four parameters: makespan, wall clock time, execution time and cost. As per the multi-objective criteria, the MOIGA algorithm can get a scheduled task.

The following are the important contributions to task scheduling research.

• •

  A Multi-objective Improved Genetic Algorithm (MOIGA) is proposed to address multi-objective task scheduling problems, which rapidly converges to a relatively close solution.

• •

  This method satisfies multi-objective optimization in the cloud by minimizing the makespan, wall clock time, execution time and cost required to execute it.

• •

  Simulation is performed using CloudSim tool.

Literature Review

Sourabh Budhiraja et al. (2012) introduced a modified genetic algorithm (MOGA) for scheduling jobs on the CPU using a cloud platform. The objective of fitness is to maximize the possibility of achieving both cost-effective and timely solutions. By adjusting the CPU's input variables, the expected result is a decreased CPU cost and improved system performance. By increasing the population and maintaining a steady population size, the MOGA can meet demand at a lower cost and improve scheduling.