Prediction of properties of starch matrix foam composites by radial basis function artificial neural network based on improved particle swarm optimization

A prediction model of starch matrix foam composites by radial basis function artificial neural network (RBF ANN) based on chaotic self-adaptive particle swarm optimization algorithm with population entropy diversity and convergence divergence strategy was established. The input variables of this model included ethylene-vinyl acetate (EVA)/starch mass ratio, glycerin content and NaHCO₃ content, and the output variables were tensile strength and rebound rate. The results show that the proposed model has a good performance. The root mean square error of prediction and correlation coefficient are 0.0160 and 0.9890, respectively. The prediction results show that the tensile strength of starch matrix foam composites reduces slowly with the increase of glycerin content, and it reduces firstly and then increases with the increase of NaHCO₃ content. The rebound rate increases with the increase of glycerin content, and it increases firstly and then decreases with the increase of NaHCO₃ content in the starch matrix composites.