基于BP神经网络及模糊推理的温度预警模型研究

Study on Logistics Temperature Prediction Model Based on BP Neural Network and Fuzzy Inference

主要研究建立疫苗冷链物流运输过程中的温度监控预警模型,通过优化的BP神经网络算法进行温度的预测,并采用模糊推理进行有效的决策预警,旨在把冷链物流运输中可能产生的损失降到最低;仿真测试阶段通过建构一个隐藏层神经元为13个的优化BP神经网络,在Matlab中进行有效性仿真,训练回归统计R值接近于1,且得出期望输出与实际值相差无几;模糊推理系统采用trapmf隶属函数,通过仿真的规则曲面表明该规则对输入有良好的判断。

This paper is to set up a temperature monitoring and early warning model in the vaccine cold chain logistics process. Through the temperature prediction by using the BP neural network algorithm, and the effective decision--making early warning by using the fuzzy inference. To reduce the possible losses in the cold chain logistics process. In simulation stage, construct a 13 hidden neurons＇ BP neural network. Through MATLAB simulation, training regression R very close to 1, and the desired output is really close on actual value too; the fuzzy inference system use the trapmf function. The simulation of regular surfaces show that the rules have good judgment.