Water Demand Forecasting Using Multi-layer Perceptron and Radial Basis Functions

Abstract

Reliable and effective management of an existing water supply entity requires both long-term and short-term water demand forecasts. Conventionally, demographic and statistical models have been employed in modeling water demand forecasts. The technique of artificial neural networks has been proposed as an efficient tool for modeling and forecasting in recent years. The primary objective of this study is to investigate artificial neural networks for forecasting both short-term and long-term water demand in the Gauteng Province, in the Republic of South Africa. Neural network architectures used in this paper are the multi-layer perceptron (MLP) and the radial basis function (RBF). It was observed that the RBF converges to a solution faster than the MLP and it is the most accurate and the most reliable tool in terms of processing large amounts of non-linear, non-parametric data in this investigation.