基于ARIMA-LSTM的物流网络分拣中心货量预测模型探究

打开文本图片集

中图分类号：0211 文献标志码：A 文章编号：2095-2945（2025）13-0046-05

Abstract：Withthevigorousdevelopmentofe-commercelogisticsnetworks，enhancinglogisticstransportationficiencyand reducinglaborcostshavebecomecoreelementsforthelogisticsindustrytostrengthenitscompetitivenessThisstudyfocuseson thecargovolumeforecastingfortransportroutesinthesortingcentersofe-commercelogisticsnetworks，aiming toacurately depictthedailyandevenhourlyfluctuationsincargovolumeoverthenext3Odaysthroughin-depthminingofhistoricalcargo volumedata.Specificall，wefirstconducted meticulouspreprocessingofdailycargovolumedatafromthepastfourmonthsand hourlycargovolumedatafromthepast3Odaysfrom57sortingcenters.Onthisbasis，weconstructedanAutoregresive IntegratedMovingAverage（ARIMA）modeltopredictthedailycargovolumeprofileoverthenext3Odays.Furthermore，we introducedaLong Short-TermMemory（LSTM）neuralnetworkmodeltoachieveprecisepredictionsof hourlycargovolumedata for the next 30 days.

Keywords： logistics network; transportation route;cargo volume profile; ARIMA model; LSTM model

现代物流是国民经济的核心，对经济发展至关重要，能促进市场繁荣、推动高质量发展。（剩余4099字）