Modellingstanddensityindexofpopulusshelterbeltswithann

For more effective farmland shelter-forest stand density simulation technology, in order to improve the management level, based on the BoAiXian in henan province from farmland shelter-forest belt for research object, by measuring method, investigated tree 135 (the) average DBH and stand behind a number density, the application of artificial neural network method was used to study the farmland shelter-forest stand density index model, reached the following results and conclusions:(1) the average DBH as inputs, and stand density for output variables, the establishment of farmland shelter-forest stand density model for a single input and single output of BP artificial neural network model.(2) with the actual data of the built model for training and analysis, the best model is a structure for 1:2:1 BP artificial neural network model, the training 0.000314 instructions for the error, precision for 98.18%.(3) and Reineke stand density index fitting accuracy of model results show that, compared with the artificial neural network model Reineke model are obtained good effect, fitting average fitting accuracy can reach more than 98%, but the simulation result shows that BP artificial neural network model conforms to the forest stand average diameter with tight density of the "S" shape variation, can simulate young stage and standard of smooth change processes in diameter. Therefore, the artificial neural network is a more effective stand density index modeling technique.(4) BP artificial neural network, which can be used as optimization technology improvement of genetic algorithm.(5) this is the characteristic of artificial neural network, the first will stand density index for farmland shelter-forest simulation, and more effective than the classical Reineke model of the simulation result.