Improved differential search algorithm based dynamic resource allocation approach for cloud application

作     者：Ma, Anxiang Gao, Yan Huang, Liping Zhang, Bin 

作者机构：Northeastern Univ Sch Comp Sci & Engn Shenyang Liaoning Peoples R China 

出 版 物：《NEURAL COMPUTING & APPLICATIONS》 (神经网络计算与应用)

年 卷 期：2019年第31卷第8期

页      面：3431-3442页

核心收录：

学科分类：08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：National Natural Science Foundation Program of China [61572116  61572117  61502089] 

主　　题：Cloud computing Differential search algorithm Dynamic resource allocation Deep belief networks 

摘      要：The performance of a service-based system (SBS) in a cloud environment may not satisfy service-level agreement (SLA) constraints when the system load changes. To improve the profits of resource providers and satisfy the global SLA, it is necessary to dynamically allocate proper resource for SBS based on the forecasted system load. By analyzing the complex workflow of the SBS, this paper proposes improved differential search algorithm-based dynamic resource allocation approach which adopts an active mechanism to respond to the change of system load so as to ensure the timely response to change. The dynamic resource allocation model based on costs optimization and SLA constraint is then proposed. The improved differential search algorithm is designed to solve the dynamic resource allocation model. This paper proposes a load forecasting approach based on deep belief networks (DBNs) in order to accurately forecast the load to support dynamic resource allocation. Experimental results show that the approach performs well in terms of the quality of the solution compared with other related approaches.