ANALYSIS OF GROUNDWATER LEVEL FLUCTUATION IN A PLAIN AREA USING GENETIC ALGORITHMS AND AN ARTIFICIAL NEURAL NETWORK

Abstract

This paper reports on a research study that investigated a robust artificial neural network (ANN) and linear combination enhanced by genetic algorithms (LC-GA) technique for analyzing groundwater level (GL) in a plain area of the Saitama prefecture in Japan. The back propagion algorithm is used in ANN model. The input sets were selected by employing an analytical technique, the cross-correlation of monthly GL. The major objective of this study was to develop a reliable groundwater level fluctuation analysis system by means of GL prediction, which have different fluctuation patterns in a plain area generating trend forecasts for the forthcoming GL monitoring and management. In general, the LC-GA model gives better prediction in testing period than the ANN model even though it has out range from training data. It was found that by inserting one time lag gives better prediction results for ANN and LC-GA models.