Computer Engineering and Applications ›› 2015, Vol. 51 ›› Issue (1): 49-53.

Previous Articles     Next Articles

Data storage engine based on memory-mapped file for evolutionary algorithms

JIANG Sanyi, DAI Zhenzhen, LI Yang, ZHOU Aimin   

  1. Department of Computer Science & Technology, East China Normal University, Shanghai 200241, China
  • Online:2015-01-01 Published:2015-01-06

基于内存映射文件的进化算法数据存储引擎

姜三义,代真真,李  阳,周爱民   

  1. 华东师范大学 计算机科学技术系,上海 200241

Abstract: In order to observe and analyze the execution of Evolutionary Algorithms(EAs), large volumes of data generated during algorithm execution are always stored in disk files. An embedded data storage engine, named as the Evolutionary Algorithm Database(EADB), provides simple yet flexible data storage interfaces for evolutionary algorithms, and facilitates fast large data storage by using memory-mapped files. Compared to storage methods using traditional file I/O and general database storage engine, EADB is tremendous faster.

Key words: evolutionary algorithm, memory-mapped file, data storage engine, file I/O

摘要: 为了观察和分析进化算法的执行情况,往往需要将算法执行过程中产生的大量数据存储在磁盘文件中。用于进化算法的嵌入式数据存储引擎EADB(Evolutionary Algorithm Database)提供了简便灵活的数据存储接口,通过使用内存映射文件技术来实现数据的快速和大量存储。相较于传统文件I/O存储方式和一般的通用数据存储引擎,EADB大大加快了存储速度。

关键词: 进化算法, 内存映射文件, 数据存储引擎, 文件I/O