Primena neuro-fazi metodologije u vodoprivredi

Abstract

U radu je izložena osnovna koncepcija dve nove metodologije za tretiranje neizvesnosti: teorije rasplinutih (fuzzy) skupova i neuralnih mreža. Razmatrene su mogućnosti primene ovih metodologija za rešavanje zadataka planiranja i upravljanja u vodoprivredi. Problemi neodređenosti i nepreciznosti, koji se često sreću pri rešavanju hidrotehničkih problema, mogu se prevazići primenom i ovih pristupa. Posebno su istaknute prednosti korišćenja kombinovanih neuro - fazi sistema, koji ispoljavaju najbolje osobine obe metodologije. Rad je metodološkog karaktera, tako da prokazane fazi veličine za klase vodotoka imaju ilustrativan značaj.

The basic concept of two methodologies fuzzy set theory and neural networks is presented in the paper. The possibility of their application in solving problems of planning in water resources management are reviewed. In water resources planning for systems with uncertainties, approach of soft modeling is more suitable then classical set theory. Advantages of hybrid neuro - fuzzy systems, that capture the best of both methodologies, is specially emphasized in the paper. The fuzzy sets for water quality classification has illustrative importance, because of the methodological character of the paper.