电子学报 ›› 2022, Vol. 50 ›› Issue (2): 314-325.DOI: 10.12263/DZXB.20201352

• 学术论文 • 上一篇    下一篇

基于镜像选择序优化的MART算法

李志博, 李清宝, 兰明敬, 孙剑帆   

  1. 中国人民解放军战略支援部队信息工程大学,河南 郑州 450001
  • 收稿日期:2020-11-30 修回日期:2021-10-29 出版日期:2022-02-25 发布日期:2022-02-25
  • 通讯作者: 兰明敬
  • 作者简介:李志博 女,1982年生,副教授,工学博士.主要研究方向为软件测试、大数据、知识图谱. E-mail: lizhibo1019@163.com
    李清宝 男,1967年生,教授,博导,工学博士.主要研究方向为信息安全、软件测试. E-mail: 13653716702@139.com
    兰明敬(通讯作者) 男,1982年生,副教授,硕导,工学博士.主要研究方向为知识图谱、大数据测试.E-mail: lanmingjing@126.com
  • 基金资助:
    国家自然科学基金(61402525);国家社会科学基金(15AJG012);国家“核高基”科技重大专项(2013JH00103)

MART Algorithm Based on Mirror Selection Order Optimization

LI Zhi-bo, LI Qing-bao, LAN Ming-jing, SUN Jian-fan   

  1. PLA Strategic Support Force Information Engineering University,Zhengzhou,Henan 450001,China
  • Received:2020-11-30 Revised:2021-10-29 Online:2022-02-25 Published:2022-02-25
  • Contact: LAN Ming-jing

摘要:

镜像自适应随机测试(Mirror Adaptive Random Testing,MART)算法将输入空间划分为多个不相交的相等子域,源域中使用自适应随机测试(Adaptive Random Testing,ART)算法生成测试数据,剩余子域中使用镜像函数生成镜像测试数据.镜像策略的引入减少了ART算法的计算开销,但是算法的检错有效性也随之降低.通过研究MART算法的特征,分析如何提升MART算法的检错有效性.针对镜像函数将源测试数据镜像到各子域时的镜像顺序,对比分析镜像选择序与镜像函数对MART算法的影响,本文提出了基于镜像受限选择序的MART算法,通过约束镜像选择序,使镜像测试数据分布更均匀.在仿真实验与实例实验结果中均显示,针对镜像策略中镜像选择序的优化,提高了MART算法的检错有效性.

关键词: 软件测试, 随机测试, 自适应随机测试, 镜像自适应随机测试, 镜像选择序

Abstract:

The input space is divided into several disjoint equal subdomains by mirror adaptive random testing(MART). The adaptive random testing(ART) algorithm is used to generate the test data in the source domain, and the mirror function is used to generate the mirror test data in the remaining subdomains. The introduction of mirror strategy reduces the computational cost of ART algorithm, but the effectiveness of failure detection is also reduced. By studying the characteristics of MART algorithm, this paper analyzes how to improve the effectiveness of MART algorithm. In this paper, we compare and analyze the influence of mirror selection order and mirror function on MART algorithm, and propose an algorithm based on mirror limited selection order, which makes the test data more evenly distributed by restricting the mirror selection order. The results of simulation and empirical experiments show that the optimization of the mirror selection order in the mirror strategy improves the effectiveness of the failure detection of the MART algorithm.

Key words: software testing, random testing, adaptive random testing, mirror adaptive random testing, mirror selection order

中图分类号: