Author(s): Xiekang Wang; Zixiang Zhang; Shuyou Cao; Duo Fang
Linked Author(s):
Keywords: Runoff; Neural networks; Genetic algorithms; Simulation
Abstract: Runoff simulation and prediction in almost watersheds is an important and essential project in water management studies, safe yield computations, some environmental disposal and design of flood control structures and so on. However, the problem is complicated because of the spatial heterogeneity exhibited by the physical and geomorphologic properties that determine a watershed’s response to precipitation events. On the other hand, the runoff time series is often obtained much easily than other affected factors such as the physical and geomorphologic parameters that are necessary to design some conceptual and dynamic forecasting model for several kinds of watersheds. Fortunately, It is common and practical tool for us to forecast some problems by means of analyzing time series. In this study we adopt the neural network model that has strongly ability to predict the non-linear function relationship to analyze the characteristics of runoff time series using genetic algorithms. The simulation and prediction of runoff in a small watershed using the proposed approach reflect a high degree of accuracy.
Year: 2002