SimTIO：基于仿真的多智能体LLM框架用于组合式交通干预优化
原标题：SimTIO: A Simulation-Grounded Multi-Agent LLM Framework for Compositional Traffic Intervention Optimization
AI 摘要
SimTIO是一个基于仿真的多智能体大语言模型框架，用于组合式交通干预优化。该框架通过模拟未修改的SUMO场景生成瓶颈证据，由三个专家智能体基于仿真反馈选择参数细化方案，并在有限仿真预算内搜索兼容干预组合。在15个城市-种子案例中，SimTIO将Top-10瓶颈时间损失和网络范围延误降低了2.78%，在86.7%的案例中找到可行改进方案，优于基线方法。该框架将LLM引导的反馈细化与基于仿真的决策权分离，并通过基线冻结目标和无操作保护防止局部改进损害网络整体性能。
正文节选
Shuyang LiGraduate Student Civil and Environmental Engineering, Rensselaer Polytechnic Institutelis36@rpi.edu[Troy, New York, 12180][0009-0006-8588-2947] \TRBauthorRuimin KeAssistant Professor Civil and Environmental Engineering, Rensselaer Polytechnic Instituteker@rpi.edu[Troy, New York, 12180] Shuyang Li, Ruimin Ke SimTIO: A Simulation-Grounded Multi-Agent LLM Framework for Compositional Traffic Intervention Optimization 1 Abstract Objectives: Traffic analysts must translate diagnosed bottlene