发表机构
Singapore Management University; Nanyang Technological University; TikTok Singapore; Nanjing University of Information Science and Technology; National University of Singapore（新加坡管理大学; 南洋理工大学; 新加坡TikTok; 南京信息工程大学; 新加坡国立大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出SimpleEvol智能体循环框架，通过最小化人类先验，让LLM自主设计启发式算法，在组合优化问题上显著提升智能转换效率，挑战复杂AHD流水线趋势。
AI 中文摘要
大型语言模型（LLMs）已成为自动化启发式设计（AHD）的强大工具，能够迭代生成和改进启发式算法。然而，主流范式将LLMs嵌入到高度手工设计的进化框架中，作为狭窄的、固定的组件，例如交叉或变异操作。我们认为这误解了LLMs的本质。它将其视为专用工具而非通用推理器，限制了它们执行低级操作，并且未充分利用其自主性。此外，这些框架中大量的人类先验违反了“苦涩教训”原则，即随着计算规模扩展的通用方法优于手工设计的解决方案。这引出了一个关键问题：哪种AHD框架设计能最好地将更强的LLM能力转化为更好的启发式算法？为解决此问题，我们提出了用于LLM驱动的AHD框架的手工化程度（AHI）和智能转换效率（ICE）指标。在三个具有挑战性的组合优化问题上评估了十个LLM，我们获得了一个显著发现：人类先验较少的框架始终产生更高的ICE。基于这一发现，我们提出了SimpleEvol，一种用于AHD的智能体循环框架，它移除了几乎所有人类先验，允许LLM自主运行。SimpleEvol始终实现最高的ICE，通常以较大优势领先。我们的结果挑战了复杂AHD流水线的趋势，并指向一种更轻量级、更以模型为中心的替代方案，表明减少人类先验是随着模型智能扩展的更有效策略。源代码可在以下网址获取：此https URL。
英文摘要
Large language models (LLMs) have emerged as powerful tools for automated heuristic design (AHD), enabling iterative generation and refinement of heuristics. However, the dominant paradigm embeds LLMs as narrow, fixed components, such as crossover or mutation, within heavily hand-engineered evolutionary frameworks. We argue this misapprehends LLMs. It treats them as specialized tools rather than general reasoners, constrains them to low-level operations, and underutilizes their autonomy. Moreover, the extensive human priors in these frameworks violate the bitter lesson principle that general methods scaling with computation surpass hand-crafted solutions. This raises a key question: which AHD framework designs best convert stronger LLM capabilities into better heuristics? To address this, we propose metrics for LLM-driven AHD framework handcraftedness (AHI) and intelligence conversion efficiency (ICE). Evaluating ten LLMs across three challenging combinatorial optimization problems, we obtain a notable finding that frameworks with fewer human priors consistently yield higher ICE. Based on this finding, we propose SimpleEvol, an agent-loop framework for AHD which removes nearly all human priors and allows the LLM to operate autonomously. SimpleEvol consistently achieves the highest ICE, often by a large margin. Our results challenge the trend toward complex AHD pipelines and point to a lighter and more model-centric alternative, suggesting that reducing human priors is a more effective strategy to scale up with model intelligence. The source code is available at https://github.com/HenryZhu1029/SimpleEvol-Master.
CommentsAccepted at NeurIPS 2026. 47 pages, 13 figures
