基于安全漏洞学习的软件研发质量演化机制研究

This project studies the impacts of the security vulnerability characteristic learning and R&D social networks on software quality. In this research, we develop quality evolution theory based on software security vulnerability characteristic learning and embeddedness theory based on the innovation of software developers. The objective of the study is to improve software quality by applying the developed theories to deduce security vulnerability characteristic learning and optimize the R&D innovation network. We focus on three areas including: the quality classification of software suppliers, the relationship between software security vulnerability characteristic learning and software quality, and the relationship between the R&D employee social network and innovation quality. Specifically, we first utilize latent variables analysis to investigate the dynamics of software quality and learning capacity of suppliers. Second, we adopt Negative Binomial Model and Hidden Markov Model to study the law of software security vulnerability characteristic learning and quality evolution. Third, based on in-depth interviews and questionnaires, we build R&D employees’ social networks, extract network characteristics and relational variables, construct economic model of employee embeddedness features, and test theories about employee R&D network and innovation quality. This research makes a theoretical contribution by proposing the software security vulnerability characteristic learning based quality evolution theory and embeddedness theories for innovation of software developers. The practical contribution of this research lies in its ability to provide software suppliers with theoretical support and best practice guidance to build effective R&D teams and a better match between quality management and R&D management systems.

本项目针对安全漏洞特征学习及研发网络对软件质量影响的瓶颈问题，从不同的视角围绕软件供应商质量分类、软件安全漏洞特征学习与质量演化、研发员工整体社会网络与创新质量三个部分展开系统而深入的研究。项目拟采用潜变量分析方法对供应商的软件质量与学习能力进行分类探讨；利用负二项式模型与隐马尔科夫模型寻找供应商软件漏洞特征学习与质量演化规律，推演不同类型软件供应商漏洞学习与质量演化理论；基于深度访谈与问卷建立供应商的研发员工社会网络，提炼研发网络结构与关系变量，构建员工嵌入性特征的经济学模型并检验员工研发网络与创新质量理论；最后，围绕着创新质量提出供应商的漏洞学习机制及员工研发创新网络的优化策略，推动软件质量的改善。本研究在理论上可形成基于软件漏洞特征学习的质量演化理论，丰富和发展软件研发员工嵌入性与创新理论；在实践上为软件供应商组建高效研发团队，推进质量管理与研发管理系统的匹配提供理论指导与最佳实践。

本项目针对安全漏洞特征学习及研发网络对软件质量影响的瓶颈问题，从不同的视角围绕软件供应商质量分类、软件安全漏洞特征学习与质量演化、研发员工整体社会网络与创新质量、IT采纳与企业绩效四个部分展开系统而深入的研究。项目采用潜变量分析方法对供应商的软件质量与学习能力进行分类探讨，研究发现专有软件研发社群和开源软件研发社群的漏洞特征学习存在差异；项目利用负二项式模型与隐马尔科夫模型寻找供应商软件漏洞特征学习与质量演化规律，推演不同类型软件供应商漏洞学习与质量演化理论，由此提出专有软件研发社群和开源软件研发社群防范不同软件漏洞风险的策略；基于深度访谈与问卷建立软件供应商的研发员工社会网络，提炼研发网络结构与关系变量，构建员工嵌入性特征的经济学模型并检验员工研发网络与创新质量理论；围绕着创新质量提出供应商的漏洞学习机制及员工研发创新网络的优化策略，推动软件质量的改善；最后，针对企业信息技术采纳与企业绩效的关系进行了深入讨论，虽然软件漏洞存在风险，但是企业广泛的IT采纳可以通过与企业业务能力的融合带来价值创造，而高质量的软件产品可以降低企业的IT应用采纳风险。本研究在理论上可形成基于软件漏洞特征学习的质量演化理论，研究结论丰富和发展软件研发员工嵌入性与创新理论；研究还应用一个具体软件研发案例探讨了漏洞学习与质量的关系，在实践上为软件供应商组建高效研发团队，推进质量管理与研发管理系统的匹配提供了最佳实践。