arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2601.22896cs.AI
收藏

基于博弈论的LLM启发式发现共演化

Game-Theoretic Co-Evolution for LLM-Based Heuristic Discovery

C2DL, Institute of Automation, Chinese Academy of Sciences(C2DL,自动化研究所,中国科学院);School of Artificial Intelligence, University of Chinese Academy of Sciences(人工智能学院,中国科学院大学);Tsinghua University(清华大学)

Xinyi Ke, Kai Li, Junliang Xing, Yifan Zhang, Jian Cheng

AI总结:

本文提出ASRO框架,通过博弈论方法实现求解器与实例生成器的共演化,提升启发式发现的泛化能力和鲁棒性。

摘要:

大型语言模型(LLMs)已推动了自动启发式发现(AHD)的快速进展,但现有大多数方法主要受限于对固定实例分布的静态评估,导致在分布偏移下可能过拟合和泛化能力差。我们提出了算法空间响应预言机(ASRO),一种基于博弈论的框架,将启发式发现重新表述为求解器和实例生成器之间的程序级共演化。ASRO将它们的交互建模为双人零和博弈,双方都维护不断增长的策略池,并通过基于LLM的最佳响应预言机迭代扩展这些池,从而用自适应的、自生成的课程替代静态评估。在多个组合优化领域中,ASRO在相同程序搜索机制基础上构建的静态训练AHD基线中始终表现更优,实现了在多样化和非分布实例上的显著改进的泛化能力和鲁棒性。

英文摘要:

Large language models (LLMs) have enabled rapid progress in automatic heuristic discovery (AHD), yet most existing methods are predominantly limited by static evaluation against fixed instance distributions, leading to potential overfitting and poor generalization under distributional shifts. We propose Algorithm Space Response Oracles (ASRO), a game-theoretic framework that reframes heuristic discovery as a program level co-evolution between solver and instance generator. ASRO models their interaction as a two-player zero-sum game, maintains growing strategy pools on both sides, and iteratively expands them via LLM-based best-response oracles against mixed opponent meta-strategies, thereby replacing static evaluation with an adaptive, self-generated curriculum. Across multiple combinatorial optimization domains, ASRO consistently outperforms static-training AHD baselines built on the same program search mechanisms, achieving substantially improved generalization and robustness on diverse and out-of-distribution instances.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑