用规则框架让小模型协作解难题，准确率从3%升至95%
Know the Ropes: A Heuristic Strategy for LLM-based Multi-Agent System Design
- 将领域知识转为分层任务分解规则，逐级拆解问题
- 在背包问题上，小模型协作准确率达95%（原3%）
- 适合想用小模型做复杂任务的开发者
单智能体大模型面临上下文有限、角色过载和领域迁移脆弱等瓶颈。传统多智能体方案虽缓解部分问题，却带来任务分解不当、契约模糊和验证开销大等新挑战。为此，我们提出知其绳（KtR）框架，将领域先验转化为算法蓝图层级结构，任务被递归拆分为有类型、受控器调度的子任务，每个子任务通过零样本或最小增强（如思维链、微调、自检）解决。基于无免费午餐定理，KtR以通用提示替代盲目搜索，实现结构化分解。在背包问题（3-8项）上，三个GPT-4o-mini代理经修复单一瓶颈后，5项实例准确率从3%提升至95%。在更难的任务分配问题（6-15任务）中，六代理o3-mini蓝图在规模10内达到100%，13-15规模达84%，相较零样本的11%显著提升。算法感知的分解结合精准增强，使小型模型成为可靠合作者，无需依赖不断膨胀的巨型模型。
原文摘要 · Abstract (English)
Single-agent LLMs hit hard limits--finite context, role overload, and brittle domain transfer. Conventional multi-agent fixes soften those edges yet expose fresh pains: ill-posed decompositions, fuzzy contracts, and verification overhead that blunts the gains. We therefore present Know-The-Ropes (KtR), a framework that converts domain priors into an algorithmic blueprint hierarchy, in which tasks are recursively split into typed, controller-mediated subtasks, each solved zero-shot or with the lightest viable boost (e.g., chain-of-thought, micro-tune, self-check). Grounded in the No-Free-Lunch theorem, KtR trades the chase for a universal prompt for disciplined decomposition. On the Knapsack problem (3-8 items), three GPT-4o-mini agents raise accuracy from 3% zero-shot to 95% on size-5 instances after patching a single bottleneck agent. On the tougher Task-Assignment problem (6-15 jobs), a six-agent o3-mini blueprint hits 100% up to size 10 and 84% on sizes 13-15, versus 11% zero-shot. Algorithm-aware decomposition plus targeted augmentation thus turns modest models into reliable collaborators--no ever-larger monoliths required.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。