Flexible Automated Software Test Generation Using AI Planning Technology

使用 AI 规划技术灵活地自动生成软件测试

• 批准号: 9619787
• 负责人: Anneliese Andrews
• 金额: $ 15万
• 依托单位: Colorado State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1997
• 资助国家: 美国
• 起止时间: 1997-08-15 至 2000-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9619787&HistoricalAwards=false
• 关键词: Flexible; Automated; Software; Test; Generation

9619787 Software testing is a critical part of software development. A new approach to software testing based on AI Planning is investigated. The similarity of plans to test cases for some systems (e.g., command language and transaction based systems) makes the representation well suited for test generation. The representation encourages testers to think at a higher level when designing tests. To exploit these potential advantages, an automated test generator is constructed with an AI planning system at its core. This generator allows representation of testing goals and testing criteria to support a variety of testing tasks. The project's contribution to Software Testing is to extend how testing goals are articulated. Its contribution to AI planning is developing new search strategies for efficient planning, and determining how to generate goals that conform to tester's intent, and evaluating the efficacy of using AI planners during software testing. Tasks include 1) improve planning efficiency for generating large tests; 2) include representation of test focus and various test criteria to make the approach flexible; 3) represent high level testing goals to reduce tedium for testers; 4) evaluate the efficacy of the new approach to identify strengths and weaknesses of using AI Planning for testing. ***

9619787软件测试是软件开发的关键部分。研究了一种基于人工智能规划的软件测试新方法。一些系统（例如，命令语言和基于事务的系统）的测试用例计划的相似性使得这种表示非常适合于测试生成。这种表示鼓励测试人员在设计测试时进行更高层次的思考。为了利用这些潜在的优势，构建了一个以人工智能规划系统为核心的自动化测试生成器。该生成器允许表示测试目标和测试标准，以支持各种测试任务。项目对软件测试的贡献是扩展了测试目标的表达方式。它对AI计划的贡献是为有效的计划开发新的搜索策略，并确定如何生成符合测试人员意图的目标，以及评估在软件测试期间使用AI计划器的有效性。任务包括：1)提高生成大型测试的计划效率；2)包括测试焦点和各种测试标准的表示，使方法灵活；3)表示高水平的测试目标，以减少测试人员的单调乏味；4)评估新方法的有效性，以确定使用AI规划进行测试的优势和劣势。＊＊＊