杉木人工林神经网络密度效应模型

Neural Network Density Effect Model of Cunninghamia lanceolata Plantation

基于人工神经网络具有逼近任意非线性映射的特性,将其应用于建立杉木人工林密度效应模型,并用免疫进化算法优化求解人工神经网络参数。结果表明:基于人工神经网络及免疫进化算法的杉木人工林密度效应模型,建模方法科学合理,林分蓄积量和平均胸径预测误差小,一定程度上优于以往传统的林分密度效应模型,在林分密度控制中有推广应用价值。

The characteristics of arbitrary non-linear mapping approximation of the artificial neural network was applied to establish the density effect model of Cunninghamia lanceolata plantation, and the parameters of the artificial neural network were obtained by using immune evolutionary algorithm. The simulation results showed that the prediction errors of the stand volume and mean DBH calculated by the artificial neural network based density effect model of C. lanceolata plantation were small, indicating that the modeling method was scientific and reasonable. This model was proved to be better than the traditional density effect models to a certain extent, it would be worthy of being extended to the stand density control for C. lanceolata plantations.