| 摘要: |
| 目的 实现基于反馈人工神经网络算法的冷链物流包装方案定制的应用终端研究与开发。方法 通过实验，建立各种影响因素下的冷藏物流要求的保温箱保温性能数据集;总结经典反馈神经网络算法的固有局限性，在此基础上提出一种更稳健、更高效的算法——SAHId-Elman，基于该算法设计冷链物流包装方案定制的人机交互界面。结果 所提出的SAHId-Elman模型在判定系数R2、均方误差(Mean Squared Error，MSE)和平均绝对误差(Mean Absolute Error，MAE)等评价指标上均优于其他3种模型，分别达到0.999 88、0.006 38、0.056 31;所设计的人机交互界面的应用终端，无需专业技术背景即可操作，单次运行时间仅为26 s，预估保温时间约12.86 h，与物理实验中的约12.92 h相比，误差仅为0.06 h，预估准确率达99.53%。结论 基于优化的SAHId-Elman算法设计的应用终端能够准确预估保温时间，表明它在冷链物流包装方案制定应用中具有可行性和可靠性。 |
| 关键词: 冷链物流 保温箱 反馈人工神经网络 保温性能 应用终端 |
| DOI：10.19554/j.cnki.1001-3563.2025.07.018 |
| 分类号: |
| 基金项目:国家新闻出版署智能与绿色柔版印刷重点实验室招标课题(ZBKT202405);高水平大学科建设医工交叉创新项目(10-22-309-501) |
|
| Application Terminal with Feedback Neural Network Algorithm for Cold Chain Packaging Solution Customization |
|
ZENG Taiying, ZHOU Sanqi, YANG Jiawen, ZHANG Chenguang
|
|
(University of Shanghai for Science and Technology, Shanghai 200093, China)
|
| Abstract: |
| The work aims to achieve research and development of an application terminal for customized cold chain logistics packaging solutions based on the feedback artificial neural network algorithm. Through experiments, a dataset of insulation performance of refrigerated logistics requirements under various influencing factors was established. The inherent limitations of classic feedback neural network algorithms were summarized. On this basis, a more robust and efficient algorithm - SAHId-Elman was proposed. Based on this algorithm, a human-computer interaction interface for customized cold chain logistics packaging solutions was designed. The proposed SAHId-Elman model outperformed the other three models in terms of evaluation indicators such as coefficient of determination R2, mean square error MSE, and mean absolute error MAE, reaching 0.999 88, 0.006 38, and 0.056 31 respectively; The application terminal of the designed human-computer interaction interface could be operated without a professional technical background, with a single run time of only 26 seconds, and the predicted insulation time was 12.86 hours, with an error of only 0.06 hours compared with the 12.92 hours in the physical experiment, and the prediction accuracy rate was 99.53%. In conclusion, the application terminal designed based on the optimized SAHId-Elman algorithm can accurately predict insulation time, indicating its feasibility and reliability in the application of cold chain logistics packaging solution formulation. |
| Key words: cold chain logistics insulation box feedback neural network insulation performance application terminal |
渝公网安备 50010702501716号