Multiobjective design of water distribution networks using modified NSGA-II algorithm

Abstract Optimally designed water distribution networks (WDNs) make engineers’ tasks difficult due to various challenges like non-linearity between head-loss and flow, commercially available distinct diameters, combinatorial, nondeterministic polynomial-time hard problems a large number of decision variables. This paper develops new hybrid NSGA-II algorithm augmented with random multi-point crossover operator local search denoted by RLNSGA-II design the multiobjective WDN. The efficiency proposed (RLNSGA-II) is tested on three benchmark problems, namely New York, Hanoi Balerma networks. results obtained are compared best-known algorithms in literature. have shown that has found better converged distributed solutions for all representative considered literature consistently evidently when approximation published. Furthermore, as complexity WDN increases, its advantages over other become more significant.