抄録
River water level of prediction Real-time plays a very important part due to making early warning about a flooding that should be decrease a flooding disaster in small river basins within the Tokyo metropolis. It is a subject that security of urban river is to estimate an early warning about the flooding on real time, that because of a run-off phenomenon is a very short time in urban river basin.
River water level is predicted due to forecast a rainfall by using distributed rainfall-runoff model by inputting. The parameter recently could get from the data of 5 floods, an error less than 0.3m on water level is reproduced to accurate rainfall using by Genetic Algorithms, that have the ability to be excellent in untying an optimization problem are used to decide parameters of the model.
