基于遗传BP神经网络的城市物流需求量预测  

Urban Logistics Demand Forecast Based On Genetic and BP Neural Network Algorithms

作　　者：阮清方[1,2] 缪立新[1] 杨朋[1] 任丽[1] 

机构地区：[1]清华大学深圳研究生院现代物流研究中心,深圳518055 [2]深圳技师学院,深圳518045

出　　处：《武汉理工大学学报（交通科学与工程版）》2011年第6期1276-1279,共4页Journal of Wuhan University of Technology（Transportation Science & Engineering）

摘　　要：结合遗传算法与BP神经网络算法预测城市物流需求量,通过算例对比证明了遗传BP神经网络算法在预测的精度与收敛速度上均优于单一算法.基于物流业的广泛性提出采用3种物流需求量作为网络的输出指标,提高了物流需求量预测的广度与可信度,并提出了一种连续预测未来数年物流需求量的方法以便于运用于实际决策之中.This article predicts demand for urban logistics on genetic algorithm and BP neural network algorithms,proving that the genetic BP neural network algorithms is better than a single algorithm in the forecast accuracy and convergence rate.Based on the extensive logistics industry,it proposes to adopt three logistics demand as the network＇s output indicators,improves the breadth and reliability of prediction,and puts forward a method of continuous forecasting in order to be used in actual decision-making.

关 键 词：遗传算法 BP神经网络算法 预测 物流 需求量 

分 类 号：U11[交通运输工程]