Study On Forest Management Decision-making Models Based On Artificial Neural Network～*

In this treatise, the Huangfengqiao forest farm in Youxian county is taken as the research object. The applications of artificial neural network in forest management decision-making are discussed, including forest resources prediction and site quality evaluation. During the course of forest management, it is important to grasp the status in quo in time and predict the development trend of forest resources for macro-management decision. The BP models for predicting the forest resources, including wooded area prediction, stand volume prediction and age class structure prediction, are developed in this treatise. The adaptability tests for the developed forest resources prediction models indicate that it is feasible for the method of the artificial neural networks to predict forest resources dynamic varieties, and can satisfy the precision of forestry production. The forest resources prediction models will provide us with powerful scientific basis for scientific management of forest resources. According to the practical condition in Huangfengqiao forest farm, on the basis of site classification, site index curve of Chinese fir, quantitative site index and variable density stand volume for BP models are developed respectively with the method of the artificial neural networks. The test method of regression equation adaptability is used to test the simulative effects of models, respectively. The testing results indicate all models, which can constitute the basis of site classification and evaluation system for Huangfengqiao forest farm, are practicable and no system deviation exist.