Optimizing the design and operation of urban water distribution networks (WDNs) is a complex, nonlinear problem. The optimization of WDNs can be performed for the pumping schedule, the location and elevation of reservoirs, the physical characteristics of pipes, and the placement of pressure-reducing valves, among other objectives. This study applies the anarchic society optimization (ASO) algorithm to find the optimal location and elevation of auxiliary tanks in urban water networks. The ASO is validated with mathematical benchmark functions, and is implemented to determine the location and elevation of auxiliary tanks in two urban water networks. The fuzzy reliability index for urban water network ranges between 74 and 79%, which is close to the global optima. The ASO exhibited better performance optimizing the reliability of WDNs than the genetic algorithm.