An EDA-GA Hybrid Algorithm for Multi-Objective Task Scheduling in Cloud Computing

Shanchen Pang,Wenhao Li,Xun Wang,Hua He,Zhiguang Shan

doi:10.1109/access.2019.2946216

Abstract

As one of the hot issues in cloud computing, task scheduling is an important way to meet user needs and achieve multiple goals. With the increasing number of cloud users and growing demand for cloud computing, how to reduce the task completion time and improve the system load balancing ability have attracted increasing interest from academia and industry in recent years. To meet the two aforementioned goals, this paper develops an EDA-GA hybrid scheduling algorithm based on EDA (estimation of distribution algorithm) and GA (genetic algorithm). First, the probability model and sampling method of EDA are used to generate a certain scale of feasible solutions. Second, the crossover and mutation operations of GA are used to expand the search range of solutions. Finally, the optimal scheduling strategy for assigning tasks to virtual machines is realized. This algorithm has advantages of fast convergence speed and strong search ability. The algorithm proposed in this paper is compared with EDA and GA via the CloudSim simulation experiment platform. The experimental results show that the EDA-GA hybrid algorithm can effectively reduce the task completion time and improve the load balancing ability.

Highlights

In recent years, cloud computing has become a hot research topic, and it is widely used in telecommunications, manufacturing, education and scientific research [1], [2]
This paper proposes a multi-objective task scheduling model that defines the demands of the tasks for virtual machines in detail
The results showed that the algorithm was superior to GA and DE in terms of quality of service and load balancing

Summary

Introduction

Cloud computing has become a hot research topic, and it is widely used in telecommunications, manufacturing, education and scientific research [1], [2]. Storage clouds [3] provide secure data storage, backup and recording services, which provide great convenience for users. Educational clouds [4] can virtualize various types of hardware education resources and transmit them to the internet system, providing a convenient information platform for education departments, teachers and students. Resources such as hardware, software and platforms are provided as services with the ‘‘pay-asyou-go’’ model. Users need to pay for only the services or resources they need without having to purchase hardware infrastructure. The current studies focus on virtualization, resource management, cloud security, green computing, task scheduling, and so forth. As cloud computing services rapidly grow, how to effectively schedule tasks to computational

Objectives

Methods

Findings

Conclusion