Self-Organized Agents: A LLM Multi-Agent Framework toward Ultra Large-Scale Code Generation and Optimization

arXiv:2404.02183v1 Announce Type: cross
Abstract: Recent advancements in automatic code generation using large language model (LLM) agent have brought us closer to the future of automated software development. However, existing single-agent approaches face limitations in generating and improving large-scale, complex codebases due to constraints in context length. To tackle this challenge, we propose Self-Organized multi-Agent framework (SoA), a novel multi-agent framework that enables the scalable and efficient generation and optimization of large-scale code. In SoA, self-organized agents operate independently to …

abstract agent agents arxiv automated code code generation constraints context cs.ai cs.cl cs.lg cs.ma cs.se development face framework future however improving language language model large language large language model limitations llm multi-agent optimization scale software software development type