Neural network based wood property mapping modeling using particle swarm optimization

As an organic whole, there are unknown nonlinear relationships existing in the different parameters of wood. This paper is aimed to solve the complex nonlinear relationship of wood parameters. Maoershan larch is selected for the test material. A neural network model is adopted with the density of wood ring and moisture content as the model inputs, wood vertical elastic modulus as the output. Particle swarm optimization is used to optimize the model. Modeling and Simulation results show that the optimization technique based on PSO modeling method is feasible and effective, with high generalization ability of the model and forecast accuracy.