造林地规划的神经网模型及其生产力预测

An Artificial Neural Network Model and Prediction of Productivity for Afforestation Program

系统地介绍了神经网络满意约束模型在造林规划上的应用情况 ,该模型简单方便 ,可实现树种优化安排。最终在计算机上模拟了这个过程 ,结果表明利用该模型收敛速度较快 ,并具有良好的适应性。满意约束模型是根据树种的生物学特性与立地条件 ,做到“适地适树”与“适树适地”为原则的一种模型。同时 ,可以由已有的立地条件信息 ,采用数量化生产力模型 ,进行预测生产力。通过实例表明 ,采用这种模型进行造林规划 ,其地位级基本上都能达到Ⅰ级。

The application of a satisfied restriction model of artificial neural network to afforestation program was discussed. The model, which is simple and convenient, can realize the optimization of species arrangements. It showed that the model has good adaptability with a quick convergence speed by computer simulation. The satisfied restriction model is based on the principles of 'suitable specie matches with suitable site' and 'suitable site matches with suitable specie'. According to obtained data about site, the productivity of a stand can be predicted by a mathematic productivity model. The site class can reach A grade basically by using this model in afforestation program.