An extensive literature review of the subject reveals that no general algorithm exists. The choice of method depends on the characteristics of the water system concerned, on the availability of data and on the objectives and constraints specified. Yeh (1985) presented the state of the art and discussed in detail various techniques, mainly for reservoir operations, including linear programming (LP), dynamic programming (DP), non-linear programming (NLP). Combinations of the above methods, along with in-depth analyses, and the merits and limitations of each of them, are also been reported in the literature.
When problems of optimal water resource management include objectives that are difficult to describe due to subjectivity or uncertainty, the principles of fuzzy logic offer a viable approach. The concepts and operational algorithms are available in literature (Zadeh and Kacprzyk, 1992; Parent and Duckstein, 1993; Russel and Campbell, 1996).
More recently neural network approaches (NNAs) have been proposed for solving conjunctive use problems. This technique has been shown to be a powerful tool for both forecasting and optimization purposes, especially when the underlying data – generating processes are uncertain or unknown. Neural network approaches have been used by French et al. (1992) for rainfall forecasting, and by Karunanithi et al (1994) for river flow prediction. Rizzo and Dougherty (1994) used a neural kriging network to characterize aquifer properties, and more recently Raman and Chandramouli (1996) proposed a neural network model for reservoir operating policies.