OPTIMAL DESIGN OF WATER DISTRIBUTION NETWORKS USING FUZZY ANALYTICAL HIERARCHY PROCESS

Water distribution networks are one of the main infrastructuresof a country. Because of high costs of designand construction, minimizing the cost as a matter ofscientic research is important. Minimizing the costshould be associated with achieving the required pressureat any point,reliability of the network. So,the integrated urban water management can be consideredas a multi-objective problem. This research proposeda new method for the optimal design by comparingdierent algorithms to optimize water distributionnetworks as a multi-objective problem using AnalyticHierarchy Process (AHP),Fuzzy Analytical HierarchyProcess (FAHP) in order to minimize the costsand maximize the three indicators of reliability indexincludes resiliency index (In), total surplus head index(It),the minimum surplus head index (Im). Thesealgorithms include genetic algorithm (GA), Honeybeemating optimization (HBMO), ant colony optimization(ACO), combinatorial optimization algorithm (AOC),Tabu search (TS), genetic algorithm linked with linearprogramming (GA-ILP), particle swarm optimization(PSO), state transition algorithm (STA),mockopen tree topology.AHP is the multi attribute decision method, which usespairwise comparisons with numerical judgments. Whencomparing two elements, the uncertain numerical ratiois expressed in a fuzzy manner rather than an exact one.Fuzzy AHP was introduced to capture the 'fuzziness' orthe vagueness,uncertainty in the evaluation of alternatives.So the factor's weights were calculated usingboth AHP,FAHP methods.To evaluate the performance,eciency of the proposedmodel, Hanoi water distribution network was chosenas a case study. Results show that Genetic algorithm(!, 10.9031) satises cost,the three indicators ofreliability criteria with both compared methods of AHPand FAHP are better than other algorithms,the lastranks belong to algorithms in which at least one nodewith pressure is less than that with the minimum allowablepressure.