CAREER: Software Reliability and Security Risk Assessment: Modeling and Algorithms

美国和世界经济体软件行业的持续增长是全球繁荣与稳定的推动力，其形式是数万亿美元的收入和数百万有价值的职业。计算的普及增加了社会对可靠和安全的软件的依赖，以确保国家安全和系统安全。生产可靠和安全的软件是困难的，因为它越来越复杂，这给许多大公司和政府组织带来了严峻的挑战。评估软件可靠性和安全性的定量模型可以帮助组织按时和在预算内交付软件系统。与软件可靠性和安全风险管理科学相关的智力优点包括：指导软件获取和集成的多阶段模型，结合数值技术以在不影响稳定性的情况下提高模型拟合过程的性能的多阶段算法，以及自动应用软件可靠性方法的开源应用程序，加强国际研究界的合作，促进研究人员和实践者之间的教育和对话。开源应用程序将众包模型和算法的贡献，为参与者建立一个公共平台来实施结果、混合建模和进行算法研究。多阶段模型将适用于从概率上用标准方法无法准确表征的复杂科学现象和工程活动。多阶段算法将适用于统计学应用的许多领域的模型拟合。开放源码工具具有将过去40年的结果与未来的发现相结合的潜力，为如何利用国际研究界提供了见解。开放式架构将以一种目标受众可以立即使用的格式，为协作和快速研究过渡提供一种媒介。开源项目通过GitHub在线服务(https://github.com/LanceFiondella/srt.core)，)的Education ator账户托管，包括所开发模型的样本输入数据集、开源工具的源代码和教育材料。主要研究成果将发表在开放获取期刊上，并可在项目网站(https://sasdlc.org/lab/projects/srt.html).)上查阅只要GitHub教育者账户仍然是一项免费的在线服务，项目库就会得到维护。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Consistent growth in the software sector of the US and world economies is a driving force for global prosperity and stability in the form of trillions of dollars in revenue and millions of rewarding careers. The spread of computing has increased society's dependence on reliable and secure software to ensure national security and system safety. Producing reliable and secure software is difficult because of its growing complexity, which poses serious challenges to many large companies and government organizations. Quantitative models to assess software reliability and security can assist organizations to deliver software systems on time and within budget.Intellectual merits related to the science of software reliability and security risk management include: multi-stage models to guide software acquisition and integration, multi-phase algorithms that combine numerical techniques to enhance the performance of model-fitting procedures without compromising stability, and an open source application to automatically apply software reliability methods, strengthen collaboration within the international research community, and promote education and dialogue between the research and practitioner communities. The open source application will crowdsource the contribution of models and algorithms, establishing a common platform for participants to implement results, hybridize modeling, and conduct algorithmic research.Multi-stage models will be applicable to scientific phenomenon and engineering activities that are too complex to accurately characterize with standard methods from probability. Multi-phase algorithms will be applicable to model-fitting in the many areas to which statistics is applied. The open source tool possesses the potential to hybridize results from the past 40 years as well as future discoveries, providing insight on how to leverage the international research community. The open architecture will provide a medium for collaboration and rapid research transition in a format that is immediately usable by the intended audience. The open source project is hosted through an Educator account of the GitHub online service (https://github.com/LanceFiondella/srt.core), including sample input data sets for the models developed, the source code of the open source tool, and educational materials. The major research findings will be published in open access journals and also made accessible on the project web site (https://sasdlc.org/lab/projects/srt.html). The project repository will be maintained as long as GitHub educator accounts remain a free online service.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.