Articles | Volume 10, issue 4
Hydrol. Earth Syst. Sci., 10, 603–608, 2006
https://doi.org/10.5194/hess-10-603-2006
Hydrol. Earth Syst. Sci., 10, 603–608, 2006
https://doi.org/10.5194/hess-10-603-2006

  07 Sep 2006

07 Sep 2006

Optimising training data for ANNs with Genetic Algorithms

R. G. Kamp1,2 and H. H. G. Savenije1 R. G. Kamp and H. H. G. Savenije
  • 1Section of Water Resources, Delft University of Technology, Delft, The Netherlands
  • 2MX.Systems B.V., Rijswijk, The Netherlands

Abstract. Artificial Neural Networks (ANNs) have proved to be good modelling tools in hydrology for rainfall-runoff modelling and hydraulic flow modelling. Representative datasets are necessary for the training phase in which the ANN learns the model's input-output relations. Good and representative training data is not always available. In this publication Genetic Algorithms (GA) are used to optimise training datasets. The approach is tested with an existing hydraulic model in The Netherlands. An initial trainnig dataset is used for training the ANN. After optimisation with a GA of the training dataset the ANN produced more accurate model results.