Stochastic Models And Its Development In Flood Area Composition

This paper analyses the rainstorm's forming cause of this researching field, and use the stochastic theory to research stochastic models' employing and development in flood area composition.What is called stochastic modeling is the process of building the stochastic model that can predict the future hydrological condition and using model through statistic test to obtain the modeling series on the basis of statistic characteristic of the field data and stochastic changing law. This paper builds three models, they are typical disaggregating model, multi-dimension auto-aggressive model and artificial neural network model, which the former two models are parameter model and the latter is nonparametric model. This paper first researched two routine models in the flood area composition, then in the base of this two modeling and made an inquiry into three stochastic models' applying in the flood area composition stochastic modeling.First, this paper researched the applying of the typical disaggregating model and multi-dimension autogressive model in the flood area composition. Because the parametric model has done some simplicity and assuming to the hydrological phenomenon, what in fact is very complicated, so between two fact emerges some deviation. To compensate these faults,the author researches a new model-BP network series model that is a nonparametric model. It has not done some predict to researching object's total density, so it can deal with all kinds of data, be suitable of more problems, be better firmness and estimate the true density.Through stochastic simulation in the area composition of the basin of lancing river, the three models can content the need of the produce practice. So, all of stochastic models are feasible and satisfied.In order to describe the complex character of the flood area composition and discover its natural laws, some stochastic models have been examined in this paper. The research work has perfected and enriched the stochastic simulating approaches of stochastic models and broadened stochastic hydrology study. This research is very successful.