A flood forecasting neural network model with genetic algorithm

Abstract

It will be useful to attain a quick and accurate flood forecasting, particularly in a flood-prone region. The accomplishment of this objective can have far reaching significance by extending the lead time for issuing disaster warnings and furnishing ample time for citizens in vulnerable areas to take appropriate action, such as evacuation. In this paper, a novel hybrid model based on recent artificial intelligence technology, namely, a genetic algorithm (GA)-based artificial neural network (ANN), is employed for flood forecasting. As a case study, the model is applied to a prototype channel reach of the Yangtze River in China. Water levels at the downstream station, Han-Kou, are forecasted on the basis of water levels with lead times at the upstream station, Luo-Shan. An empirical linear regression model, a conventional ANN model and a GA model are used as the benchmarks for comparison of performances. The results reveal that the hybrid GA-based ANN algorithm, under cautious treatment to avoid over-fitting, is able to produce better accuracy in performance, although at the expense of additional modelling parameters and possibly slightly longer computation time.