arXiv cs.MA· Ziyao Zeng, Hao Wang, Chen Liu, Youheng Yao, Jingcheng Ni, Tianyu Liu, Xiatao Sun, Fengyu Yang, Chenyu You, Xiaofeng Liu, Daniel Rakita, Ronald R. Coifman, Yuval Kluger, Zhiwen Fan·· 13 小时前AI 评分34
RuleSmith：多智能体 LLM 实现游戏规则自动平衡
RuleSmith: Multi-Agent LLMs for Automated Game Balancing
AI 导读
RuleSmith 将多智能体 LLM 自博弈与基于代理模型的优化结合，在参数化规则空间中自动搜索游戏平衡配置。概念验证游戏 CivMini 含两个不对称阵营和 12 个可调参数，三次独立优化运行在 200 与 300 局评估中均达到接近均势的胜率。该方法无需任务特定策略训练即可输出可解释的数值规则调整，NTBEA 与 TPE 实验表明同一自博弈接口可搭配不同搜索组件。
正文 · 原文
Authors:Ziyao Zeng, Hao Wang, Chen Liu, Youheng Yao, Jingcheng Ni, Tianyu Liu, Xiatao Sun, Fengyu Yang, Chenyu You, Xiaofeng Liu, Daniel Rakita, Ronald R. Coifman, Yuval Kluger, Zhiwen Fan
Abstract:Game balancing is a longstanding challenge requiring repeated playtesting, expert intuition, and extensive manual tuning. We introduce RuleSmith, a framework that couples multi-agent LLM self-play with surrogate-based optimization over a parameterized rule space. As a proof of concept, we build CivMini, a simplified civilization-style game with two asymmetric factions governed by 12 tunable parameters. LLM agents interpret textual rulebooks and game states to propose actions that are validated by the game engine, yielding noisy estimates of faction win rates. Our primary implementation uses Gaussian-process Bayesian optimization with Expected-Improvement-based adaptive sampling to allocate more evaluation games to promising candidates. RuleSmith consistently finds near-balanced configurations: three independent optimization runs reach near-even empirical win rates in independent 200- and 300-game evaluations, and additional experiments with NTBEA and TPE show that the same self-play interface supports different search components. Cross-model evaluations and small human studies provide preliminary evidence of transfer across player populations. RuleSmith produces interpretable numerical rule adjustments without task-specific policy training. Project and code: this https URL
| Comments: | Accepted at NeurIPS 2026 |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Science and Game Theory (cs.GT); Multiagent Systems (cs.MA) |
| Cite as: | arXiv:2602.06232 [cs.LG] |
| (or arXiv:2602.06232v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2602.06232 arXiv-issued DOI via DataCite |
Submission history
From: Ziyao Zeng [view email]
[v1]
Thu, 5 Feb 2026 22:19:44 UTC (11,031 KB)
[v2]
Wed, 21 Oct 2026 03:12:56 UTC (11,050 KB)
来源：arXiv cs.MA · arxiv.org