Artificial neural network based generalized storage–yield–reliability models using the Levenberg–Marquardt algorithm A.J. Adeloye a, * , A. De Munari b a School of the Built Environment, Heriot-Watt University, Edinburgh EH14 4AS, UK b Via Dolarici 18, 25015 Desenzano del Garda BS, Italy Received 1 March 2005; revised 18 August 2005; accepted 27 October 2005 Abstract Generalised storage–yield–reliability models are developed using multi-layer perceptrons artificial neural networks (ANNs), trained using the Levenberg–Marquardt algorithm. These ANNs provide, for the first time, generalised models for simultaneously predicting within-year and over-year storage capacities, given the yield, reliability and readily obtainable streamflow statistics. The training, validation and testing of the models used time series data from 18 streams located in different parts of the world, which were carefully selected so that they nearly cover the range of flow variability observed in world streams. The performance of the models was very good. Further comparison of the ANN models with existing regression models revealed that the latter are marginally better; however, given that the regression models require the over-year capacity to be known a priori, the ANN models are more generic and should be preferred. q 2006 Elsevier B.V. All rights reserved. Keywords: Artificial neural networks; Storage–yield–reliability; Levenberg–Marquardt; Sequent peak algorithm; Over-year capacity; Within- year capacity 1. Introduction Reservoir planning involves the determination of storage capacity of the reservoir, which will satisfy the demand with an acceptable level of reliability. This information can be provided by developing the storage–yield–reliability (S–Y–R) relationship for the site under investigation. The development of the S–Y–R function is generally achieved by carrying out a sequential analysis of time series data using one of a variety of traditional techniques such as behaviour simulation and the Sequent Peak Algorithm (see McMahon and Adeloye, 2005). However, more recently, other techniques have been developed which by-pass the sequential analysis of time series data. Because they do not require the sequential analysis of time series data, these generalised techniques, as they are commonly known (see Vogel and Stedinger, 1987) are much faster to implement, which makes them very useful during preliminary Journal of Hydrology 362 (2006) 215–230 www.elsevier.com/locate/jhydrol 0022-1694/$ - see front matter q 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jhydrol.2005.10.033 * Corresponding author. Fax: C44 131 451 4617. E-mail addresses: a.j.adeloye@hw.ac.uk (A.J. Adeloye), annal- isa.demunari@virgilio.it (A.D. Munari).