Scholarlink

<p>Software reliability plays a pivotal role in ensuring the quality of developed software systems, particularly in safety–critical applications and has become a primary concern for both software developers and users. This growing concern has triggered the need to characterise reliability quantitatively and precisely predict software performance over time. One such framework is Software Reliability Growth Modelling (SRGM), which has emerged as a specialised field that aids in characterising software error correction and detection behaviour over time. It makes use of mathematical and statistical techniques for the assessment of the software reliability measures, such as mean time to failure, the mean number of errors left and the type of unresolved error. Various SRGMs have been introduced in the literature; each adds a unique perspective on characterising software reliability. This paper aims to analyse 40&#xa0;years of research conducted in SRGMs using bibliometric methodology. The application of the bibliometric technique reveals the contributions of different research constituents, including authors, institutions, and journals, to this research field. By analysing scientific production from 215 distinct sources indexed in Web of Science and Scopus, this analysis explores the relationship between research constituents using different science mapping techniques. An important finding is the clear evolutionary trajectory of the field, which has transitioned from foundational, statistics-based models towards a diverse application of AI. Our analysis shows a significant surge in AI-related research post-2012, now focused on Deep Learning and evolutionary algorithms. The comprehensive analysis of this research field offers useful perspectives into the evolution of the SRGM field traditional debugging to modern AI-driven approaches, offering clear perspectives on the field’s past, present, and future directions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

From debugging to AI-driven paradigms: a bibliometric analysis of four decades of software reliability growth models

  • 期刊论文

  • 发表日期：

    2025年11月10日

查看全文
  • Reena Kasana,
  • Jagvinder Singh,
  • Meela Rani,
  • Balraj Singh

摘要

Software reliability plays a pivotal role in ensuring the quality of developed software systems, particularly in safety–critical applications and has become a primary concern for both software developers and users. This growing concern has triggered the need to characterise reliability quantitatively and precisely predict software performance over time. One such framework is Software Reliability Growth Modelling (SRGM), which has emerged as a specialised field that aids in characterising software error correction and detection behaviour over time. It makes use of mathematical and statistical techniques for the assessment of the software reliability measures, such as mean time to failure, the mean number of errors left and the type of unresolved error. Various SRGMs have been introduced in the literature; each adds a unique perspective on characterising software reliability. This paper aims to analyse 40 years of research conducted in SRGMs using bibliometric methodology. The application of the bibliometric technique reveals the contributions of different research constituents, including authors, institutions, and journals, to this research field. By analysing scientific production from 215 distinct sources indexed in Web of Science and Scopus, this analysis explores the relationship between research constituents using different science mapping techniques. An important finding is the clear evolutionary trajectory of the field, which has transitioned from foundational, statistics-based models towards a diverse application of AI. Our analysis shows a significant surge in AI-related research post-2012, now focused on Deep Learning and evolutionary algorithms. The comprehensive analysis of this research field offers useful perspectives into the evolution of the SRGM field traditional debugging to modern AI-driven approaches, offering clear perspectives on the field’s past, present, and future directions.