摘要
为准确预测感潮河段船舶交通流,提出一种复合潮汐信息的感潮河段船舶交通流滚动预测模型。选择多变量非线性核函数灰色预测模型KGM(1,N)作为基础模型,并将潮汐信息作为右端项核函数输入信息;针对KGM(1,N)模型存在的不足,采用插值系数法进行背景值优化;运用粒子群优化(particle swarm optimization,PSO)算法确定核函数所需的高斯核参数、修正参数和最优背景值系数;在输入数据有限的情况下,采用实时滚动预测方法,保证模型充分利用新信息。经上海港南槽航道九段警戒区上游的船舶交通流实例验证,所提出的模型具有较高的预测精度。
In order to accurately predict ship traffic flow in tidal reaches,a rolling prediction model of ship traffic flow in tidal reaches combined with tidal information is proposed.The kernel function-based multivariate nonlinear grey model KGM(1,N)is selected as the basic model,and the tidal information is used as the input information for the kernel function at the right-hand side;for the shortcomings of KGM(1,N)model,the interpolation coefficient method is used to optimize the background value;the particle swarm optimization(PSO)algorithm is used to determine the Gaussian kernel parameter,the correction parameter and the optimal background value coefficient of the kernel function;under the condition of limited input data,the real-time rolling prediction method is used to ensure that the model makes full use of new information.The proposed model is of high prediction accuracy as verified by the example of ship traffic flow in the upstream section of the Jiuduan precautionary zone in Nancao Channel of Shanghai Port.
作者
齐绪存
黄常海
沈佳
娄乃元
QI Xucun;HUANG Changhai;SHEN Jia;LOU Naiyuan(Merchant Marine College,Shanghai Maritime University,Shanghai 201306,China)
出处
《上海海事大学学报》
北大核心
2022年第4期37-43,82,共8页
Journal of Shanghai Maritime University
基金
国家自然科学基金(51909156)
上海市科技创新行动计划(18DZ1206300)
上海市软科学研究项目(20692108700,21692193000)。
关键词
感潮河段
船舶交通流
交通流预测
灰色预测模型
背景值优化
tidal reach
ship traffic flow
traffic flow prediction
grey model
background value optimization
作者简介
齐绪存(1998-),男,江苏徐州人,硕士研究生,研究方向为交通信息工程及控制,(E-mail)1401080078@qq.com;黄常海(1987-),男,山东滕州人,讲师,博士,研究方向为水上交通态势感知、监测、预警等,(E-mail)changhai406@126.com。
