广东省水产预制菜冷链物流需求分析及预测

徐超毅，胡望敏

安徽理工大学经济与管理学院，安徽 淮南  232001

摘要：为预测广东省水产预制菜冷链物流的市场需求，从区域经济发展水平、市场供需水平、交通运输水平、冷链技术水平4个方面构建水产预制菜冷链物流需求评价指标，采用灰色关联分析法研究影响水产预制菜冷链物流需求变化的主要因素，分别采用灰色模型（grey model，GM）中的一阶一元微分方程GM（1，1）与长短时记忆（long short-term memory，LSTM）神经网络对比分析2015—2021年广东省水产预制菜冷链物流需求。结果表明：影响广东省水产预制菜冷链物流需求发展的主要影响因素为货物周转量和冷链冷藏水平；GM(1，1)、LSTM神经网络预测结果的平均相对误差分别为2.68%、0.22%，后者的预测准确度明显优于前者；采用LSTM神经网络预测2022—2024年广东省水产预制菜冷链物流需求，广东省水产预制菜需求呈上升趋势，预计2024年将达到509.09万t。广东省应立足冷链基础设施建设，确保贮藏、运输过程中水产预制菜温度稳定，加强水产预制菜食品监督，保证食品质量安全，不断促进水产预制菜冷链产业的发展。

关键词：水产预制菜；冷链物流需求；GM（1，1）；LSTM神经网络；灰色关联分析

Analysis and forecast of cold chain logistics demand for aquatic pre-made dishes in Guangdong Province

XU Chaoyi, HU Wangmin

School of Economics and Management, Anhui University of Science and Technology, Anhui 232001, China

Abstract: In order to predict the market demand for aquatic pre-made dishes in Guangdong Province, an evaluation index of cold chain logistics demand for aquatic pre-made dishes is constructed from four aspects: regional economic development level, market supply and demand level, transportation level, and cold chain technology level. The main factors affecting the change of cold chain logistics demand for aquatic pre-made dishes are studied by gray correlation analysis method. The first-order unitary differential equation GM(1,1) in the gray model (GM) and long short-term memory (LSTM) neural network are used to compare and analyze the cold chain logistics demand for aquatic pre-made dishes in Guangdong Province from 2015 to 2021. The results showed that the main influencing factors affecting the development of cold chain logistics demand for aquatic pre-made dishes in Guangdong Province are cargo turnover and cold chain refrigeration level. The average relative errors of GM(1,1) and LSTM neural network are 2.68% and 0.22%, respectively, and the prediction accuracy of the latter is significantly better than that of the former. Using LSTM neural network to predict the cold chain logistics demand forcargo in Guangdong Province from 2022 to 2024, the demand for cargo in Guangdong Province is on the rise, and it is expected to reach 509.09 million tons in 2024. Guangdong Province should focus on cold chain infrastructure building, ensure stable temperature of aquatic pre-made dishes during storage and transportation, strengthen food supervision of aquatic pre-made dishes, ensure food quality and safety, and continuously promote the development of the aquatic pre-made dishes cold chain industry.

Keywords: aquatic pre-made dish; cold chain logistics demand; GM (1,1); LSTM neural network; grey relational analysis