Research on Short-distance Route Planning for Vehicle Loading Based on Hybrid Prediction

View at publisher Summarise with VIRAT
Authors: Yueyang Hu, Yehang Linghu, Ting Zhou, Jingjing Chen, Shiyun Nie
Cited by: 0
References: 7

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

A hybrid predictive model was developed that fused LSTM and Holt-Winters techniques to analyze load time series at 10-minute granularity. LSTM networks can effectively capture nonlinear modes. The results show that prediction errors of less than 5% for specific routes, such as "Site 3 -Site 83 - 060," help improve scheduling. The multi-objective scheduling optimization model aims to improve vehicle turnover, package volume per vehicle, and reduce costs. Greedy algorithm handles node strings and vehicle assignments, resulting in an 18% increase in vehicle turnover, 932 pieces per vehicle, and a 12% cost reduction over traditional methods in 72 line combination tests. For unstable routes like Site 3 - Station 83 - 140, dynamic adjustments in the node string reduce the deadhead by 25%.

Fields of Study & Subjects

Vehicle Routing Optimization Methods Advanced Manufacturing and Logistics Optimization Transport and Logistics Innovations

Publication details

Journal / source
IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
Pages
375–380
Publication date
28 January 2026
Language
EN