Modelling Daily Streamflow using Genetic Algorithm

Abstract

This paper presents the result of modelling daily streamflow for a catchment area in Malaysia by using rainfall and water level data. The data used in this paper are based on the data recorded at the Sungai Pahang basin obtained from Department of Irrigation and Drainage (DID) and this paper applies an approach of Genetic Algorithm model. Genetic programming software called Discipulus is used for modelling the daily streamflow. In this paper, the data of the rainfall and water level are extracted and filtered before the data are transformed into Training, Validation and Applied Data for optimization process of Discipulus. The daily streamflow model that generated from Discipulus is compared to actual streamflow data. The major aim of the research is to forecast daily streamflow based on rainfall and water level data, to evaluate the performance and reliability of the time series model for daily streamflow forecasting where evaluation of model performance is based on Genetic Algorithm modelling by using Discipulus. The result shows that genetic programming able to predict reliable model of daily streamflow for Sungai Pahang basin and it is recommended that more catchment areas to be modeled using genetic programming in order to achieve better result interpretation.