船舶交通流组合预测方法研究  

Research on Combination Forecasting Method for Ship Traffic Flow

作　　者：刘敬贤[1] 张涛[1] 刘文[2] 

机构地区：[1]武汉理工大学航运学院,湖北武汉430063 [2]武汉理工大学理学院,湖北武汉430070

出　　处：《中国航海》2009年第3期80-84,共5页Navigation of China

摘　　要：船舶交通流预测的研究能为港口设计及船舶通航安全管理提供基础性数据。基于港口特征、船舶行为和船舶交通流的历史数据,建立了一种改进的船舶交通流变权组合预测模型,较好地解决了现有船舶交通流预测算法中存在的预测精度不高、依赖于经验的不足。以天津港主航道连续9年的历史观测数据为例进行验证,结果表明,改进的组合预测模型能够减小单一模型预测中存在的不确定性,从而提高了整个预测系统的精度及其稳定性。Prediction research on ship traffic flow can provide basic data for harbor layout design and management of ship navigation. A modified combination forecasting model is set up for forecasting ship traffic flow, based on harbor characteristics, ship behavior and historical data of ship traffic flow to remedy shortcomings of present ship traffic flow prediction systems such as low precision and dependence on experience. Through analyzing the nine-year ship flux data of Tianjin harbor, results show that this modified combination forecasting model can reduce uncertainty in prediction so as to enhance precision and stability of the whole forecasting system.

关 键 词：水路运输 船舶交通流 组合预测模型 港口特征 船舶行为 

分 类 号：U692.37[交通运输工程—港口、海岸及近海工程]