作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2021, Vol. 47 ›› Issue (2): 246-253,260. doi: 10.19678/j.issn.1000-3428.0057688

• 体系结构与软件技术 • 上一篇    下一篇

动态可靠性约束的多阶段测试资源分配研究

占德志1, 张国富1,2,3, 苏兆品1,2,3, 岳峰1,2   

  1. 1. 合肥工业大学 计算机与信息学院, 合肥 230601;
    2. 合肥工业大学 工业安全与应急技术安徽省重点实验室, 合肥 230601;
    3. 安全关键工业测控技术教育部工程研究中心, 合肥 230601
  • 收稿日期:2020-03-12 修回日期:2020-04-14 出版日期:2021-02-15 发布日期:2020-04-21
  • 作者简介:占德志(1993-),男,硕士研究生,主研方向为软件工程;张国富,教授、博士;苏兆品,副教授、博士;岳峰,副研究员、博士。
  • 基金资助:
    国家自然科学基金(61573125);教育部人文社会科学研究青年基金项目(19YJC870021,18YJC870025);中国工程院咨询研究重点项目(2020-XZ-3);中央高校基本科研业务费专项资金(PA2019GDQT0008,PA2019GDPK0072)。

Research on Multi-Stage Testing Resource Allocation with Dynamic Reliability Constraints

ZHAN Dezhi1, ZHANG Guofu1,2,3, SU Zhaopin1,2,3, YUE Feng1,2   

  1. 1. School of Computer Science and Information Engineering, Hefei University of Technology, Hefei 230601, China;
    2. Anhui Province Key Laboratory of Industry Safety and Emergency Technology, Hefei University of Technology, Hefei 230601, China;
    3. Engineering Research Center of Safety Critical Industry Measure and Control Technology, Ministry of Education, Hefei 230601, China
  • Received:2020-03-12 Revised:2020-04-14 Online:2021-02-15 Published:2020-04-21

摘要: 为满足测试资源分配过程中用户对软件可靠性的需求,构建一种动态可靠性约束的多阶段多目标测试资源分配模型DRC-MSMOTRA。从理论上分析不同阶段满足可靠性约束的测试时间下限并设计相应的种群初始化策略,结合参数估计、加权归一化方法和多目标差分进化,提出一种动态可靠性约束的多阶段多目标测试资源分配算法MS-DRC-GDE3。实验结果表明,与MSMOTRA模型相比,DRC-MSMOTRA模型在2种不同规模的软件系统上所获解的覆盖值分别提高约62和59个百分点,与MS-GDE3算法相比,MS-DRC-GDE3算法在2种软件系统上所获解的覆盖值分别提高约69和80个百分点,即所提模型和算法能够根据用户对可靠性的需求来为用户提供更多更优的测试资源分配方案。

关键词: 软件可靠性, 测试资源分配, 动态可靠性约束, 加权归一化, 多目标差分进化

Abstract: To meet the user's requirements for software reliability in testing resource allocation,this paper constructs a multi-stage multi-objective testing resource allocation model called DRC-MSMOTRA with dynamic reliability constraints.The lower bounds of the testing time to meet the reliability constraints in different stages are theoretically analyzed and the corresponding population re-initialization strategy is designed.Then a multi-stage multi-objective testing resource allocation algorithm with dynamic reliability constraints,MS-DRC-GDE3,is developed according to parameter estimation,weighted normalization and multi-objective differential evolution.The experimental results show that compared with the MSMOTRA model,the proposed DRC-MSMOTRA model increases the coverage value of the obtained solution on two different scales of software systems by 62 and 59 percentage points respectively.Compared with the MS-GDE3 algorithm,the proposed MS-DRC-GDE3 algorithm increases the coverage value of the obtained solution on two software systems by 69 and 80 percentage points respectively,which demonstrates that the proposed model and algorithm can provide more optimized testing resource allocation schemes for users based on their requirements for reliability.

Key words: software reliability, testing resource allocation, dynamic reliability constraint, weighted normalization, multi-objective differential evolution

中图分类号: