港口吞吐量预测的动态平滑参数新模型  

A New Dynamic Smoothing Parameter Model of Port Throughput Forecast

作　　者：黄洪琼[1] 王天真[1] 

机构地区：[1]上海海事大学,上海200135

出　　处：《中国航海》2006年第4期86-89,共4页Navigation of China

基　　金：国家自然科学基金NSFC(NO.60572051);上海市教委项目:05FZ25;05FZ04

摘　　要：港口吞吐量预测是港口总体布局规划中重要的前期工作。将最优梯度算法应用于指数平滑模型,通过构造优选并自动生成的最佳平滑参数,使模型得以优化,增强了指数平滑模型对时间序列的适应能力。较好地解决了指数平滑预测中,平滑参数靠检验确定且为静态,平滑初值难以确定并导致预测偏差等问题。以上海港口为例,与其它指数平滑预测算法比较,验证了该算法的有效性和实用性。Throughput forecast is the important work in early stage of general port layout. A new exponential smoothing model with dynamic smoothing parameter and without selecting initial smoothing parameter is proposed in the paper. The algorithm is constructed to create optimal parameter for model optimization, which enables the model to adapt time sequence better so that the problems, e.g. the parameter is static and determined only by experiences, smoothing initial value is hard to be determined and may lead to forecast deviations, will be settled. Taking Shanghai port as an example, this model is compared with other exponential smooth forecast methods, its effectiveness and practicability are validated.

关 键 词：水路运输 港口吞吐量 预测 平滑参数 最优梯度 

分 类 号：U691.71[交通运输工程—港口、海岸及近海工程]