Behavioral Level Design Verifications Using Software Testing Techniques and Neural Networks

使用软件测试技术和神经网络进行行为级别设计验证

• 批准号: 9628770
• 负责人: Thomas Chen
• 金额: $ 32.43万
• 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: 1996
• 资助国家: 美国
• 起止时间: 1996-08-01 至 2000-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9628770&HistoricalAwards=false
• 关键词: Behavioral; Level; Design; Verifications; Using

A structured methodology for validating and verifying behavioral models is being explored. The methodology is adapted from a set of software test and validation criteria, which are based on control and data flow properties of the behavioral model under test. The test criteria are measured by a set of coverage metrics which indicates the types and the locations of untested portions in the model. The methodology includes an adaptive test data generation technique using neural networks. Given the untested portions in the model, effective test data are generated and identified by neural networks to exercise the untested portions. Tools which will provide designers with a quantitative confidence index at the end of their testing and verification efforts are being developed.

正在探索一种用于确认和验证行为模型的结构化方法。该方法改编自一组软件测试和验证标准，这些标准基于被测行为模型的控制和数据流属性。测试标准由一组覆盖指标来衡量，这些指标指示了模型中未测试部分的类型和位置。该方法包括使用神经网络的自适应测试数据生成技术。在给定模型中未测试部分的情况下，生成有效的测试数据并由神经网络识别以训练未测试部分。目前正在开发工具，这些工具将在设计者的测试和核查工作结束时为其提供一个量化的信心指数。