CAREER: Software Reliability Assessment for High Assurance Systems

确保高保障系统在部署之前满足其可靠性要求是至关重要的。研究的重点是将形式化验证方法和程序测试相结合。这个统一的、定量的软件可靠性评估框架的基石是贝叶斯推理的使用。软件故障的主观概率是通过正式或非正式的验证得到的。统计测试可以证实满足要求的可靠性水平，或者表明主观建立的故障概率不现实，提示需要返工和改进。验证技术和统计测试的几种组合是基于程序结构的。研究目标包括识别程序设计，使软件可靠性评估可行，并证明其适用于高保证应用。与研究主题相关的教育活动包括本科生和研究生的研究体验以及课程开发。教育活动的目标是让学生通过参与实验和案例研究来了解复杂系统的开发和保证。参赛学生组成多学科团队，包括计算机科学、计算机工程和电气工程专业的学生。

Ensuring that high assurance systems meet their reliability requirements prior to deployment is essential. The research focuses on combining the strengths of formal verification methods and program testing. The cornerstone of this unified, quantitative software reliability assessment framework is the use of Bayesian inference. Subjective probabilities of software failures are obtained through formal or informal verification. Statistical testing can either corroborate that the required reliability level is met, or it indicates that the subjectively established failure probabilities are not realistic, prompting the need for rework and improvement. Several combinations of verification techniques and statistical testing are based on program structure. Research objectives include the identification of program designs, which make software reliability assessment feasible, and demonstration of their suitability for high assurance applications. Educational activities accompanying the research theme include research experience for undergraduate and graduate students, and curriculum development. The goal of educational activities is to expose students to the development and assurance of complex systems through their participation in experiments and case studies. Participating students form multidisciplinary teams including computer science, computer engineering and electrical engineering majors.