High performance parallel evolutionary algorithm model based on MapReduce framework. (1st January 2013)

• Record Type: Journal Article
• Authors: Du, Xin; Ni, Youcong; Yao, Zhiqiang; Xiao, Ruliang; Xie, Datong
• Is Part Of: International journal of computer applications technology. Volume 46:Number 3(2013)
• Page Start: 290
• Page End: 295
• Publication Date: 2013-01-01
• Subjects: evolutionary algorithms -- EAs -- parallel evolutionary algorithms -- MapReduce; Technology -- Data processing -- Periodicals; 620.00285
• Languages: English
• ISSNs: 0952-8091

Abstract

Evolutionary algorithms (EAs) are increasingly being applied to large-scale problems. MapReduce is a powerful abstraction proposed by Google for making scalable and fault tolerant applications. However, how to design high performance parallel EA based on MapReduce (MR-PEA) is still an open problem. In this paper, a parallel evolutionary algorithm model based on MapReduce by improving traditional parallel evolutionary algorithms model is proposed. The MR-PEA model is fit for large populations and datasets, has the characteristic of high scalable and efficiency. In order to justify the effectiveness of the MR-PEA model, we proposed a parallel gene expression programming based on MapReduce (MR-GEP) used to solve symbolic regression.