用不确定性引导采样，让水网数字孪生更准更省
Conformal Prediction-Driven Adaptive Sampling for Digital Water Twins
- 基于LSTM与置信区间预测节点不确定性
- 40%采样率下需求误差降低33%-34%
- 适合实时水网监控，计算开销仅增5%-10%
面向供水管网的数字孪生系统需在传感器有限的情况下实现精准状态估计。传统均匀采样在不同节点间资源分配不均，导致浪费。本文提出一种结合LSTM预测与置信区间（CP）的自适应采样框架，通过边际置信区间实现低延迟不确定性估计，聚焦于最不确定节点。在Hanoi、Net3和CTOWN数据集上的实验表明，当采样覆盖率达40%时，需求误差比均匀采样降低33%~34%，同时保持89.4%~90.2%的实证覆盖率，计算开销仅增加5%~10%。
原文摘要 · Abstract (English)
Digital Twins (DTs) for Water Distribution Networks (WDNs) require accurate state estimation with limited sensors. Uniform sampling often wastes resources across nodes with different uncertainty. We propose an adaptive framework combining LSTM forecasting and Conformal Prediction (CP) to estimate node-wise uncertainty and focus sensing on the most uncertain points. Marginal CP is used for its low computational cost, suitable for real-time DTs. Experiments on Hanoi, Net3, and CTOWN show 33--34\% lower demand error than uniform sampling at 40\% coverage and maintain 89.4--90.2\% empirical coverage with only 5--10\% extra computation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。