notesum.ai

Published at November 27

An End-to-End Smart Predict-then-Optimize Framework for Vehicle Relocation Problems in Large-Scale Vehicle Crowd Sensing

cs.LG

math.OC

Released Date: November 27, 2024

Authors: Xinyu Wang¹, Yiyang Peng¹, Wei Ma¹

Aff.: ¹Civil and Environmental Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong SAR, China

Arxiv: http://arxiv.org/abs/2411.18432v1

Refer to caption

Grid size	Dimension	Total budget	SPO-A(Ours)		SPO-C		PTO		DON
Grid size	Dimension	Total budget	RMSE	SMAPE (%)	RMSE	SMAPE (%)	RMSE	SMAPE (%)	RMSE	SMAPE (%)
$45\times 45$	2,025	8,000	9.072	31.623	9.510	31.551	9.563	34.081	14.517	52.643
		10,000	9.647	31.552	9.541	31.457	9.551	33.987	14.517	52.643
		12000	9.589	31.587	9.593	31.558	9.554	31.523	14.517	52.643
		15,000	9.059	31.426	9.507	31.431	9.563	34.087	14.517	52.643
$68\times 68$	4,624	8,000	9.834	33.479	9.962	33.514	10.701	38.166	15.449	56.384
		10,000	9.689	32.966	9.820	33.643	10.519	36.955	15.449	56.384
		12,000	9.291	30.504	9.678	33.282	10.098	33.162	15.449	56.384
		15,000	9.752	33.277	10.096	32.478	10.547	37.495	15.449	56.384