Abstract:Through improving the error back propagation artificial neural networks, a kind of new model of back-propagation artificial neural networks based on simulated annealing algorithm was built. The new model can overcome the disadvantage of traditional BP, which are slower convergence velocity, lower accuracy and easily fall into the local optimum. Through using the new model to analyzing the rice flood loss, it was showed that there are several advantages such as fast convergence velocity, good whole optimum, and strong practicability. The new improved model can be used in some other correlative research area. (Nature and Science. 2005;3(2):22-27).

Application of ANN Model Based Simulated Annealing to Analyzing the Rice Flood Loss