Flood discharge forecasting by Genetic Programming

• Main Authors: Wei-chun Fang, 方唯鈞
• Other Authors: Chang-hsien Chen
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/40629098653411031835

碩士 === 逢甲大學 === 水利工程所 === 94 === Genetic programming is a kind of control theory that spring up in recent years. This deterministic method evolves to adapt itself to the environment of problems. When using deterministic method to infer a hydrology process, the method maybe ignore some of the drastic events. It will increase risk of the flood forecasting. This thesis proposes the classification of typhoon data. The procedure chooses different models of the model bank that according to the structure of data to simulate flooding. After this process, we apply ANFIS to correct the error of GP model base on the rainfall and discharge. The ultimate goal of this research is to reduce the risk of flood forecasting variation. The GP model performed stably and reliably in the validation tests at Wu-Xi watershed.