SwarmResearch: Orchestrating Coding Agents for Open-Ended Discovery

SwarmResearch：编排编码智能体以实现开放式探索

Abstract: Long-running coding agents such as autoresearch can persistently discover optimizations for open-ended problems. However, they tend to converge onto a single high-level approach, then proceed with low-level edits while missing other superior approaches to the problem. We hypothesize two harness-level design choices contribute to this behavior: accumulating context in a single long-running agent and only exposing a single program state to edit.

摘要： 诸如 autoresearch 之类的长周期编码智能体能够持续发现开放式问题的优化方案。然而，它们往往会收敛于单一的高层级方法，随后仅进行低层级的编辑，从而错失了解决该问题的其他更优路径。我们推测，有两个框架层面的设计选择导致了这种行为：一是在单一长周期智能体中累积上下文，二是仅暴露单一程序状态供编辑。

We introduce SwarmResearch, an orchestrator-subagent harness in which a Shepherd Agent uses global context to steer a population of Search Agents, each operating with local context in their respective git branch. On open-ended optimization tasks, SwarmResearch discovers better or comparable solutions to state-of-the-art LLM-guided evolution and multi-agent techniques on 13/15 tasks, driven by higher-level exploration.

我们引入了 SwarmResearch，这是一个由“编排者-子智能体”组成的框架。其中，“牧羊人智能体”（Shepherd Agent）利用全局上下文来引导一群“搜索智能体”（Search Agents），每个搜索智能体都在各自的 git 分支中利用局部上下文进行操作。在开放式优化任务中，得益于更高层级的探索，SwarmResearch 在 15 个任务中的 13 个任务上，发现了优于或媲美当前最先进的 LLM 引导进化技术及多智能体技术的解决方案。

Compared with fixed scaling of serial and parallel agents, SwarmResearch’s orchestrator-guided scaling discovers better-performing solutions by adapting parallelism at different search depths.

与串行和并行智能体的固定扩展方式相比，SwarmResearch 的编排引导式扩展通过在不同搜索深度调整并行度，从而发现了性能更优的解决方案。

---

Paper Details:

• Authors: Yuvraj Virk, Zack Edds, Chunqiu Steven Xia, Lingming Zhang
• Subject: Artificial Intelligence (cs.AI)
• arXiv ID: 2607.02807

论文详情：

• 作者： Yuvraj Virk, Zack Edds, Chunqiu Steven Xia, Lingming Zhang
• 学科： 人工智能 (cs.AI)
• arXiv ID: 2607.02807