Ranking solutions of multi-objective reservoir operation optimization models using multi-criteria decision analysis

In this paper, ELECTRE-TRI method has been used for ranking the optimal solutions of Non-dominated Sorting Genetic Algorithm-II (NSGA-II) which has been developed for multi-objective operation optimization of a cascade system of reservoirs. In the multi-objective optimization model, flood control and water supply have been considered as short-term and long-term objectives, respectively. The flood control objective function has been formulated to minimize the expected flood damages in different flooding months. A short-term reservoir optimization model which provides the optimal hourly releases during floods has been developed and linked to a flood damage estimation model. The water supply objective function minimizes water deficit damages in agricultural areas. The losses associated with deficit in supplying agricultural water demands have been calculated based on crop yield responses to deficit irrigation. Selection of the most practical solution among the optimal Pareto set of solutions, which is addressed using ELECTRE-TRI method, is a challenging issue. The proposed methodology provides an opportunity for further incorporation of decision makers' preferences in ranking the NSGA-II optimal set of solutions. The results of applying the developed models to the cascade system of the Dez and Bakhtiari Reservoirs in southwest of Iran show that ELECTRE-TRI method has the ability to rank the non-dominated solutions of the multi-objective operation optimization model and provides a suitable framework for considering the preferences of decision makers (DMs). (C) 2010 Elsevier Ltd. All rights reserved.