中文

Formula One 赛车策略中 tyre 能量可解释时间序列预测

机器学习 2025-01-09 v1 人工智能

摘要

Formula One(F1)赛车策略发生在 high-pressure and fast-paced 的环境中,where split-second decisions can drastically affect race results。Pit stop decisions(即更换车辆的 tyre)和 tyre compound(硬、medium 或 soft,取决于 normal conditions)的选择是两项 core decisions。optimal 的 pit stop decisions 可以通过估计这些 compound 的 tyre degradation 来确定,tyre degradation 又可以从 each tyre 所施加的 energy,即 tyre energy 来计算。在本 work 中,我们使用由 Mercedes-AMG PETRONAS F1 队的历史 race data 包括 telemetry 来训练 deep learning models,以预测比赛中的 tyre energy。此外,我们 also fitted XGBoost,一种基于 decision tree 的 machine learning algorithm,到相同的数据集并进行比较,两者均给出 impressive performance。Furthermore,我们还整合了两种不同的 explainable AI 方法,即 feature importance 和 counterfactual explanations,以 gain insights into forecasts 背后的 reasoning。Our contributions thus result in an explainable, automated method which could assist F1 teams in optimising their race strategy。

关键词

引用

@article{arxiv.2501.04067,
  title  = {Explainable Time Series Prediction of Tyre Energy in Formula One Race Strategy},
  author = {Jamie Todd and Junqi Jiang and Aaron Russo and Steffen Winkler and Stuart Sale and Joseph McMillan and Antonio Rago},
  journal= {arXiv preprint arXiv:2501.04067},
  year   = {2025}
}

备注

9 pages, 9 figures. Copyright ACM 2025. This is the authors' version of the work. It is posted here for your personal use. Not for redistribution. The definitive Version of Record will be published in SAC 2025, http://dx.doi.org/10.1145/3672608.3707765