基于神经网络的集装箱船港口作业时间预测模型  

CONTAINER SHIP PORT OPERATION TIME PREDICTION MODEL BASED ON NEURAL NETWORK

作　　者：韩宗垒 徐斌[1] 陈佳[1] Han Zonglei;Xu Bin;Chen Jia(Liaoning Provincial Key Laboratory of Logistics and Shipping Management System Engineering,Dalian Maritime University,Dalian 116026,Liaoning,China)

机构地区：[1]大连海事大学辽宁省物流航运管理系统工程重点实验室,辽宁大连116026

出　　处：《计算机应用与软件》2021年第2期78-84,共7页Computer Applications and Software

摘　　要：集装箱船港口作业时间是制作泊位计划的一个重要依据,而集装箱船港口作业时间获取的主要来源是预测。传统预测方法是用装卸集装箱量除以岸桥装卸效率,预测精度较低,且受多种因素的影响,具有复杂的非线性特点。而神经网络在解决复杂的非线性问题方面具有很强的建模能力,所以选取神经网络建立集装箱船港口作业时间预测模型。通过真实数据对预测模型进行训练学习,用测试数据集对模型进行验证,并且与传统预测方法相对比,结果表明了该预测模型在某集装箱港口预测应用的有效性。Container ship port operation time is an important basis for the production of berth plans,and the main source of container ship port operation time is prediction.The traditional prediction method is the loading and unloading container volume divided by the shore bridge loading and unloading efficiency.The prediction accuracy of this method is low,and it is affected by many factors and has complex nonlinear characteristics.The neural network has strong modeling ability in solving complex nonlinear problems.Therefore,the neural network was selected to establish a container ship port operation time prediction model.The prediction model was trained by real data,and the model was validated by the test data set.Compared with the traditional prediction method,the effectiveness of the prediction model in a container port prediction application is demonstrated.

关 键 词：水路运输 集装箱船港口作业时间 神经网络 预测 模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]