中文

An Evaluation of Role-Based Multi-Agent Code Generation on Repository-Scale Problems

软件工程 2026-07-05 v1
作者: Benedetta Donato , Noah Hagar-Dent , Aaron Worsnop , Leonardo Mariani , Valerio Terragni
在 arXiv 上查看 ↗ PDF ↗ DOI ↗

摘要

Role-based multiagent code generation aims to make LLMs more effective on repository-scale problems, moving beyond small programming tasks. We evaluate this approach on 12 Java repositories, finding greater similarity to developer code than single LLMs, but a persistent gap from human implementations.

引用

@article{arxiv.2607.04212,
  title  = {An Evaluation of Role-Based Multi-Agent Code Generation on Repository-Scale Problems},
  author = {Benedetta Donato and Noah Hagar-Dent and Aaron Worsnop and Leonardo Mariani and Valerio Terragni},
  journal= {arXiv preprint arXiv:2607.04212},
  year   = {2026}
}

备注

8 pages, 1 figure, 2 tables. Accepted for publication in the IEEE Software Special Issue on Engineering Agentic Systems