EUVIMEDEUVIMED entdecken
Systematische ÜbersichtOpen Access · CC BY 4.0

A Systematic Literature Review on Automated Unit Test Generation Techniques, Algorithms, and Tools

Syed Mukhtar ul Hassan, Muhammad Hammad Ali, Muhammad Abbas, Mudassar Adeel Ahmed

International Journal of Innovations in Science and Technology · 19.05.2026 · 694

DOI: 10.33411/ijist/1809 ↗

Kurz gefasst

Systematische Übersicht in der Zeitschrift International Journal of Innovations in Science and Technology, online erschienen am 19.05.2026. Verfasst von Syed Mukhtar ul Hassan und 3 weiteren Autor:innen. Die Arbeit nennt 65 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The generation of automated unit tests has advanced rapidly with Natural Language Processing (NLP) and Large Language Model (LLM)-based methods. Nonetheless, the field is still disjointed in terms of techniques, algorithms, tools, and verification practices, making it difficult to determine where significant progress has been made, and what major limitations persist. This paper is a systematic literature review of 48 primary studies on automated unit test generation across four dimensions of analysis: generation techniques, underlying algorithms, tools/frameworks, and verification or post-generation improvement mechanisms. The results indicate that tools and frameworks account for the highest number of studies, with 27 (56.3%), followed by test generation techniques and algorithms with 14 (29.2%), verification and evaluation approaches with 10 (20.8%), and assertion, repair, and test suite improvement mechanisms with 8 (16.7%) studies. Evidence has been reported that the latest LLM-based and hybrid methods have shown better contextual generation and workflow integration, though these methods heavily rely on validation, repair, and quality of oracles. As an example, using mutation-guided prompting has been reported to reach a mutation score of 93.57% on synthetic buggy code and detect up to 28% more faulty human-written snippets, and industrial LLM-based test improvement has reported 75% build success, 57% reliable pass rate, 25% coverage increase, and 73% acceptance by engineers. The review in general indicates that the trend of automated unit test generation is no longer about isolated generation of tests, but rather about the generation of tests that are integrated into quality-engineering pipelines. New developments will not rely as much on the generate raw tests but on integrating generation with verification, assertion enhancement, repair, reproducible evaluation, and deploying these capabilities in industrial settings.

Zitieren

APA

Hassan, S. M. U., Ali, M. H., Abbas, M., & Ahmed, M. A. (2026). A Systematic Literature Review on Automated Unit Test Generation Techniques, Algorithms, and Tools. International Journal of Innovations in Science and Technology, 694. https://doi.org/10.33411/ijist/1809

Vancouver

Hassan SMU, Ali MH, Abbas M, Ahmed MA. A Systematic Literature Review on Automated Unit Test Generation Techniques, Algorithms, and Tools. International Journal of Innovations in Science and Technology. 2026:694. doi:10.33411/ijist/1809

BibTeX
@article{Hassan2026,
  author = {Hassan, Syed Mukhtar ul and Ali, Muhammad Hammad and Abbas, Muhammad and Ahmed, Mudassar Adeel},
  title = {A Systematic Literature Review on Automated Unit Test Generation Techniques, Algorithms, and Tools},
  journal = {International Journal of Innovations in Science and Technology},
  year = {2026},
  pages = {694},
  doi = {10.33411/ijist/1809}
}
Kostenlos starten →