Linear programming (LP)-based algorithms are the one amongst the various techniques used for the optimization of WDNs. They are quick to produce results; however, most of them, when used for design of looped WDNs, cannot move further from a local optimum solution to a global optimal solution. Further, they result in split pipe solutions which require a special connector that increases the cost of network. A solution considering a single size for each link in the network is desirable and proposed herein to obtain using an integer linear programming (ILP) approach, specifically zero–one ILP. This proposed methodology selects the most suitable size for each link, satisfying all constraints of the problem. The application of ILP is extended to obtain a solution with a single-pipe for each link using the replacement-elimination method in which an initially selected branched network is iteratively improved. The algorithm moves from one local optimum to another in search of a global optimum tree solution. The application of the methodology is shown herein with two benchmark networks. The proposed methodology is generic and can be applied to any single source gravity network.