摘要
码头集装箱装船作业时间在码头调度计划中扮演着重要角色。传统的集装箱装船时间预测方法仅仅将作业时间与工作量视为线性关系,导致预测精度较低。为了提高码头集装箱装船时间预测精度,提出了一种基于神经网络的码头集装箱装船时间预测模型。该模型以BP神经网络为基础模型,用Momentum算法优化BP神经网络。以作业过程中的多种影响因素为输入,以预测的装船作业时间为输出。通过用某大型集装箱码头半年的实际作业数据对预测模型进行训练与测试,并与线性回归和随机森林预测结果进行对比,结果表明,该模型较线性回归和随机森林具有更低的均方根误差和平均误差,总体预测误差能控制在3%以内,能较有效预测集装箱装船作业时间。
In order to improve the accuracy of container loading time prediction at the dock,a method combining neural network for container loading time prediction at the dock is proposed in this paper.This method can consider multiple influencing factors during the operation process to improve prediction accuracy.The prediction model is trained using actual operation data from a large dock over a six-month period,and then validated using test set data.Experimental results show that the model predicts the actual operation data with an error of about 3%,demonstrating that this method can effectively predict container loading operation time.
出处
《工业控制计算机》
2025年第4期85-86,89,共3页
Industrial Control Computer
基金
上海市科委科技计划项目(22511103304)资助。
