Please use this identifier to cite or link to this item:
|Title:||PREDICTION FOR NON-REVENUE AND DEMAND OF URBAN WATER USING HYBRID MODELS OF NEURAL NETWORKS AND GENETIC ALGORITHMS|
|Authors:||FARAH, BURHAN $AAUP$Palestinian|
Awad, Mohammed $AAUP$Palestinian
RUTROT, AMJAD $AAUP$Palestinian
Multilayer Perceptron NNs
Radial Basis Function NNs,
Genetic Algorithms, .
|Publisher:||Journal of Theoretical and Applied Information Technology|
|Citation:||FARAH, B., AWAD, M., & RUTROT, A. (2022). PREDICTION FOR NON-REVENUE AND DEMAND OF URBAN WATER USING HYBRID MODELS OF NEURAL NETWORKS AND GENETIC ALGORITHMS. Journal of Theoretical and Applied Information Technology, 100(21).|
|Series/Report no.:||vol 100, No. 21;6537 -6551|
|Abstract:||Palestine faces continuous struggles in maintaining the proper water supply in the water sector. Therefore, “Non-Revenue water” and supply demands are necessary to reduce the water losses and save the financial resources to strengthen the water sector. To do that we must develop the ideal water usage/loss prediction model to plan the future usage of water. This paper explores and develops AI models that could efficiently predict the water losses and water demands in Palestine, focusing mainly on Beitunia city. Different Artificial Neural Networks (ANNs) with different learning approaches had been used in this paper. The historical and extracted data, representing water supply/consumption in Beitunia are used to propose a nonlinear model. The data is input into the models of ANN and helps predict the water losses/demand in Palestine, to provide a more accurate prediction model. Three models of ANNs were used; Multilayer Perceptron NNs (MLPNNs) MLPNNs-LM, Radial Basis Function NNs (RBFNNs, newrb), and Genetic Algorithms (GAs-MLPNNs). We also used the Autoregressive integrated moving averages (ARIMA) as a linear statistical model to predict water supply using collected data from Beitunia city. The result showed that ANNs models are more efficient than the ARIMA model for the prediction of water movement. Finally, The MLPNNs-LM model results exceeded the other ANNs models in comparison|
|Appears in Collections:||Faculty & Staff Scientific Research publications|
Files in This Item:
|22Vol100No21.pdf||1.95 MB||Adobe PDF||View/Open|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.