Optimal design of water distribution system by multiobjective evolutionary methods

Abstract

Determination of pipe diameters is the most important problem in design of water supply networks. Several authors have focused on the methods capable of sizing the network considering uncertainty and other important aspects. This study presents an application of multiobjective decision making techniques using evolutionary algorithms to generate a series of nondominated solutions. The three objective functions considered here include investment costs, entropy system and system demand supply ratio. The determination of Pareto frontier employed the public domain library MOMHLib++ and a hybrid hydraulic simulator based on the method of Nielsen. This technique is found to be quite promising, the nondominated region being identified in a reasonably small number of iterations. © Springer-Verlag Berlin Heidelberg 2003.