Enhancing LLM-driven Program Repair through RAG and Test-Method Mapping 


Vol. 15,  No. 1, pp. 61-69, Jan.  2026
10.3745/TKIPS.2026.15.1.61


PDF
  Abstract

Automated Program Repair (APR) aims to automatically fix buggy programs, with recent advances in Large Language Models (LLMs) showing strong potential. However, existing methods often struggle to leverage contextual information and capture the relationships between buggy methods and failing test cases. This paper aims to enhance LLM-driven program repair by systematically integrating intelligent method-test mapping with retrieval-augmented generation to address these limitations. We present an LLM-driven framework with three key components: first, a mapping strategy that links buggy methods to relevant test cases, next, retrieval-augmented generation (RAG) that incorporates semantically similar code examples, and lastly, an iterative feedback-driven generation process. We constructed the framework using GPT-4o and evaluated it on 353 method-level bugs from Defects4J 1.2. Experimental results demonstrated that our approach efficiently repaired 181 bugs, outperforming the baseline methods through effective context utilization and guided patch generation.These results indicate that retrieval-enhanced contextual reasoning can substantially improve repair accuracy and scalability.

  Statistics
  Cite this article

[IEEE Style]

A. S. Abdinabiev, E. Jung, B. Lee, "Enhancing LLM-driven Program Repair through RAG and Test-Method Mapping," The Transactions of the Korea Information Processing Society, vol. 15, no. 1, pp. 61-69, 2026. DOI: 10.3745/TKIPS.2026.15.1.61.

[ACM Style]

Aslan Safarovich Abdinabiev, Eunseo Jung, and Byungjeong Lee. 2026. Enhancing LLM-driven Program Repair through RAG and Test-Method Mapping. The Transactions of the Korea Information Processing Society, 15, 1, (2026), 61-69. DOI: 10.3745/TKIPS.2026.15.1.61.

Forms

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)

Old Journals

Indexing

All publications of TKIPS is indexed in DOI, EBSCO, Google Scholar, Crossref, and CrossCheck.

TKIPS is also selected as the Journal for Accreditation by NRF (National Research Foundation of Korea).

Related Journals