基于RBF神经网络的水产品冷链物流需求预测研究

PREDICTION OF DEMAND FOR COLD CHAIN LOGISTICS OF AQUATIC PRODUCTS BASED ON RBF NEURAL NETWORK

关键词:  水产品冷链物流需求预测模拟仿真RBF神经网络

Key words:  aquatic products  cold chain logistics  prediction of demand  analog simulation  RBF neural network

摘要:

［目的］水产品是我国农业经济的重要组成部分，冷链物流制约着水产品的可持续发展。为实现水产品供需均衡和物流资源合理配置，有效推进水产品冷链物流行业快速发展，对水产品冷链物流需求进行预测尤为重要。［方法］选取影响水产品冷链物流需求的9个主要因素，运用RBF神经网络模型、BP神经网络模型、多元线性回归模型和GM（1,1）模型对2007—2016年我国水产品冷链物流需求分别进行模拟预测，并对预测结果进行对比分析。［结果］总体上我国居民对水产品的需求量逐年上升,2016年的需求量超过1 600万t，水产品冷链物流需求具有极大的市场空间； 在水产品冷链物流需求预测中，RBF神经网络预测结果的残差平均值、相对误差平均值和误差均方根均比BP神经网络、MLR和GM（1,1）模型的预测误差小。［结论］RBF神经网络模型对水产品冷链物流需求的预测精度显著优于其他预测方法，表现出极强的处理非线性系统的能力，说明RBF神经网络模型是预测水产品冷链物流需求的一种有效方法。

Abstract:

Aquatic products are an important part of the agricultural economy in China, while cold chain logistics impact the sustainable development of aquatic products. In order to realize the balance between supply and demand and the reasonable allocation of logistics resources, and effectively promote the rapid development of cold chain logistics, it is particularly important to predict the demand for cold chain logistics of aquatic products. By selecting 9 main influencing factors that affected the demand for cold chain logistics of aquatic products, this paper applied RBF neural network, BP neural network, the multiple linear regression and GM(1,1) to predict the demand for cold chain logistics of aquatic products between 2007 and 2016, then compared the mentioned predictions. The conclusions showed that the demand for cold chain logistics of aquatic products was increasing by and large year by year. Specifically, the demand exceeded 16 million tons in 2016. That was, cold chain logistics had great market prospects. In addition, in the forecasting of the demand for cold chain logistics of aquatic products, RBF neural network had the smaller forecasting error than that of BP neural network, MLR and GM(1,1) in the mean absolute error, the mean absolute percentage error and the root mean square error, respectively. Therefore, RBF neural network shows a strong ability to deal with nonlinear systems, and it is also superior to other forecasting methods, which indicates that RBF neural network is an effective method for forecasting the demand for cold chain logistics of aquatic products.