Predication of Landslide Based on Grey System and Evolutionary Artificial Neural Networks

摘要

Predication of landslide is very important in control of landslide disaster. Considering the monotonously increasing character of time series of the landslide displacement, a new intelligent method combining Grey System and Evolutionary Neural Network (ENN) is proposed here. In this method, based on the principles of displacement decomposition, the trend of time series is extracted by Grey System and the deviation is approximated by the new ENN proposed here. In this new ENN, the architecture and algorithm parameters can evolve simultaneously through combining modified BP algorithm and Immunized Evolutionary Programming proposed by author. This new method is applied in the Xintan landslide, and the results show that the generalization of the new method is good and it can predict the landslide very well.