The design of optimal hydraulic networks is formulated as a mathematical optimization problem. This formulation is of interest as a possible substitute for the current practice of parametric studies, as such studies are lengthy and not guaranteed to yield an optimum solution. The applicability of optimization techniques and the economical feasibility of the whole process have been investigated, and are illustrated with a sample problem. It is also shown that deficiencies in starting designs or initial network geometries may be corrected automatically by the computerized procedure, and that the effects of changes in network mode of operation or boundary conditions, or both, can easily be studied. Results may be applicable to larger hydraulic-network design problems, because the form of the mathematical expressions remains the same.