Forecasting River Flow in The Usa Using A Hybrid Metaheuristic Algorithm With Back-Propagation Algorithm

()

---

Abstract

Water resource management is a very complex and important challenge in this century; most of the political analysts consider that the world is going to face a problem of water management and reservation in the future. water management and estimation in the short and long term is an essential tool in planning, maintaining, managing and controlling the unexpected events. In this work, a Metaheuristic Algorithm which hybridizes Tabu search and Genetic algorithms with Back-propagation Algorithm (MABP) were proposed for managing, controlling and predicting water flow. The proposed algorithm used the statistical data set of the Ontonagon River near Rockland in USA as a case study. Back-propagation algorithm was used to train the Artificial Neural Network (ANN) using 550 daily data sets and was tested for other different 550 daily data sets. Metaheuristic algorithm (MA) was used to enhance and improve the obtained weights (solutions quality) from Back-propagation Algorithm (BP) by increasing and enhancing the fitness cost number. The estimated time series, ANN convergence, training and tested graphs were also explored. Key Words: Artificial Neural Networks, Back-propagation algorithm, Metaheuristic algorithm, River flow forecasting and Water resource management.
PDF

References