EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Large Language Model-Based Autonomous Software Engineering: Intelligent Code Generation, Testing, and Defect Repair

Keerthana B C, Sanjay Kumar Tuddu, Dr. Shubha Baravani, Ms. Kancherla R. Krishna Naga Surya Alekhya, Anand Kumar Mishra, Dr. Vijendra Pratap Singh

Computer and Decision Making: An International Journal · 15.11.2026 · 4 (1) · 498-511

DOI: 10.1922/comdem.v4i1.19178 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Computer and Decision Making: An International Journal, online erschienen am 15.11.2026. Verfasst von Keerthana B C und 5 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Large language models (LLMs) are reshaping the software engineering landscape by supporting intelligent automation in code generation, software testing, debugging and defect repair. The review covers the underlying concepts, the potential, and the new horizons of LLM-enabled autonomous software engineering, particularly the shift from AI-assisted software engineering to agentic and increasingly autonomous processes. It covers significant code generation paradigms, such as text-to-code, code completion, repository-level code generation, and code refinement, as well as LLM-driven test generation, assertion construction, code coverage improvement, and automated program repair. The review also addresses planning, memory, tool use, feedback, and collaborative multi-agent mechanisms that have been integrated into software engineering agents for end-to-end software development. Yet, there are many challenges to overcome such as hallucination, lack of functional correctness, security issues, limited ability to reason at repository scale, computational cost, evaluation bias, and explainability, among others. Future work needs to focus on verifying code generation, self-testing and self-repairing systems, standard test frameworks, secure agent architecture, persistent memory, and industrial testing to improve reliable and trustworthy autonomous software engineering.

Zitieren

APA

C, K. B., Tuddu, S. K., Baravani, D. S., Alekhya, M. K. R. K. N. S., Mishra, A. K., & Singh, D. V. P. (2026). Large Language Model-Based Autonomous Software Engineering: Intelligent Code Generation, Testing, and Defect Repair. Computer and Decision Making: An International Journal, 4(1), 498-511. https://doi.org/10.1922/comdem.v4i1.19178

Vancouver

C KB, Tuddu SK, Baravani DS, Alekhya MKRKNS, Mishra AK, Singh DVP. Large Language Model-Based Autonomous Software Engineering: Intelligent Code Generation, Testing, and Defect Repair. Computer and Decision Making: An International Journal. 2026;4(1):498-511. doi:10.1922/comdem.v4i1.19178

BibTeX
@article{C2026,
  author = {C, Keerthana B and Tuddu, Sanjay Kumar and Baravani, Dr. Shubha and Alekhya, Ms. Kancherla R. Krishna Naga Surya and Mishra, Anand Kumar and Singh, Dr. Vijendra Pratap},
  title = {Large Language Model-Based Autonomous Software Engineering: Intelligent Code Generation, Testing, and Defect Repair},
  journal = {Computer and Decision Making: An International Journal},
  year = {2026},
  volume = {4},
  number = {1},
  pages = {498--511},
  doi = {10.1922/comdem.v4i1.19178}
}
Kostenlos starten →