基于GM-BP的露天矿生产材料消耗分析与预测

Analysis and Prediction Material Consumption in Open-pit Mines Based on GM-BP

神经网络建模过程中隐含层节点数多采用定性、经验或者是反复实验比较结果的方式确定,一直达不到最佳的状态,引入灰色关联分析法来确定神经网络的隐含层节点数。建立GM-BP模型是结合灰色关联分析方法优化BP神经网络的隐含层结构,提高了BP网络的适应能力,使之能更有效地应用于复杂系统的建模方法。对露天矿柴油消耗进行分析和预测,可以为生产材料的合理配置提供决策支持。分析燃料柴油的主要影响因素,分别用灰色模型、多元回归分析和GM-BP模型进行比较和检验,改进后的模型具有较好的稳定性,预测准确,可为设备油耗考核和油库库存管理提供参考。

The number of hidden layer nodes in the neural network modeling process is mostly determined by quality, experience, or comparing the results of repeated experiments. The optimal state is not reached. The gray correlation analysis method is proposed to determine the number of hidden layer nodes in the neural network. The GM-BP model is established to optimize the hidden layer structure of BP neural network by combining with gray correlation analysis method, so that the adaptability of BP neural network can be improved, which makes BP neural network more effective in modeling method of complex systems. The analysis and prediction of diesel consumption in open-pit mines can provide decision support for the rational allocation of production materials. The main influencing factors of fuel diesel are analyzed, and the gray model, multiple regression analysis and the GM-BP model are used to compare and test them. The improved model has better stability and satisfactory prediction accuracy. The improved model can provide reference for equipment fuel consumption assessment and oil depot inventory management.