摘要
首先建立基于遗传BP神经网络的滚装运输市场货运量预测模型,然后采用川江滚装运输年度数据进行神经网络训练与预测。结果表明,遗传BP神经网络算法在预测的精度与收敛速度上均优于传统预测算法。预测结果为川江滚装运输企业战略规划和滚装运输竞争力的提高提供了决策支持。
In this paper, we first established the volume forecasting model of the ro- ro transportation market based on the GA- BP Neural network model, then trained the model using the practical data of the Chuanjiang city, and found that the accuracy and convergency speed of the algorithm were both superior to traditional forecasting algorithms.
出处
《物流技术》
北大核心
2014年第7期184-187,207,共5页
Logistics Technology
基金
中国博士后科学基金(2013M542269)
国家社科基金(13CGL151)
关键词
货运量
滚装运输
遗传算法
BP神经网络
川江
预测
freight volume
ro-ro transportation
genetic algorithm
BP neural network
Chuanjiang
forecasting
