基于GRA-GA-BP神经网络的港口集装箱吞吐量预测模型  

Prediction Model of Port Container Throughput Based on GRA-GA-BP Neural Network

作　　者：于少强 周钰博 陈康 肖长凯 林宇玲 YU Shaoqiang;ZHOU Yubo;CHEN Kang;XIAO Changkai;LIN Yuling(School of Maritime Economics&Management,Dalian Maritime University,Dalian 116026,China)

机构地区：[1]大连海事大学航运经济与管理学院,辽宁大连116026

出　　处：《物流技术》2022年第9期78-82,共5页Logistics Technology

摘　　要：港口集装箱吞吐量的准确预测对于港口的未来发展规划及该港口所属腹地的经济发展具有重要的指导意义。为实现输入变量的降维,提高预测精度,避免出现局部最优解,借助灰色关联分析对输入变量进行筛选和排序,引入遗传算法作为优化BP神经网络内初始权值及阈值的工具,并基于此建立模型对环渤海港口群三大代表港口2016-2019年集装箱吞吐量进行实证对比分析。最后横向比较传统BP模型和GA-BP模型的预测结果,研究结果表明:经GRA-GA算法优化BP神经网络模型相较于GA-BP神经网络模型与未经过改善的传统的BP神经网络模型其预测结果更为理想,研究结果对港口的战略规划及发展具有一定的参考价值。The paper points out that the accurate prediction of port container throughput has important guiding significance for the future development planning of a port and the economic development of the hinterland to which the port belongs. In order to reduce the dimension of input variables, improve prediction accuracy and avoid local optimization, the paper screens and ranks input variables with grey relational analysis, and introduces genetic algorithm as a tool to optimize the initial weights and thresholds in the BP neural network. On such basis, it establishes a model for empirical comparison and analysis of the container throughput of the three representative ports in the Bohai Bay Rim port cluster from 2016 to 2019. Finally, it compares the prediction result of the traditional BP model and the GA-BP model, and finds that the prediction result of the BP neural network model optimized by the GRA-GA algorithm is better than that of the GA-BP neural network model and the traditional BP neural network model without improvement, which provides reference for strategic port planning and development.

关 键 词：港口 集装箱 吞吐量预测 灰色关联度分析 遗传算法 BP神经网络 

分 类 号：U691[交通运输工程—港口、海岸及近海工程]