arXiv：cs.LG· Angel Wang, Dominique Perrault-Joncas, Alvaro Maggiar, Dean Foster, Carson Eisenach·· 3 小时前AI 评分40
从反事实模拟器推演中预测：一项 Sim2Real 评估
Forecasting from Counterfactual Simulator Rollouts: A Sim2Real Evaluation
AI 导读
研究用两个真实库存控制部署评估了模拟器训练预测模型的 Sim2Real 迁移，从模拟器保真度、零样本迁移和真实数据累积后的自适应三方面展开。模拟器训练的预测器 MAPE 低于用历史真实数据训练的相同架构，Study 1 降低 1.2-3.1 个百分点，Study 2 降低 12.5-18.7 个百分点。部署后用早期真实观测做轻量校准，误差最多再降 2.5 个百分点。
正文
Abstract:Deploying a new decision policy creates a cold-start problem for prediction models whose targets depend on the policy's actions: historical observations reflect earlier policies, while real observations under the new policy are not yet available. Simulation offers a way to address this gap by rolling out the target policy across counterfactual scenarios and using the resulting trajectories to learn how the system responds to those controls. The simulation-to-reality (Sim2Real) transfer of this simulator-trained model can then be backtested by evaluating it against real observations from past deployments. Using two real-world inventory-control deployments, we evaluate this process from three angles: simulator fidelity, zero-shot transfer to real behavior, and adaptation as real target-policy observations accumulate. The simulator-trained forecaster achieves lower point-estimate mean absolute percentage error (MAPE) than the same architecture trained on historical real data, reducing MAPE by 1.2-3.1 percentage points in Study 1 and 12.5-18.7 points in Study 2. After deployment, lightweight calibration using early real observations further reduces error by up to 2.5 percentage points. These results provide empirical evidence that simulator-generated counterfactual data can support cold-start forecasting under a new policy, and the resulting model can be further refined as real deployment data become available.
| Comments: | 15 pages, 3 figures, 9 tables |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.03662 [cs.LG] |
| (or arXiv:2610.03662v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.03662 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Angel Wang [view email]
[v1]
Fri, 2 Oct 2026 17:37:14 UTC (1,206 KB)
来源：arXiv：cs.LG · arxiv.org