A Mixed-Integer Linear Programming approach to Optimizing Water Distribution Networks

Abstract

Urban infrastructure’s backbone is water distribution networks, which ensure reliable supply of clean water to residents, institutions and industries. Due to growing cities and rising water demands, it becomes absolutely crucial to optimize those networks to be efficient, cost effective, and environmentally friendly. The application of mixed integer linear programming (MILP) for solving water distribution network problems has been the subject of this article and its strengths and weaknesses have also been discussed. Optimizing water distribution networks is a difficult problem because of the interdependency of the different factors in the pipe diameters, pump characteristics, reservoir heights, and system hydraulics. These multifaceted problems appear to be traditionally unsolved via traditional methods of optimization and thus researchers and engineers dive into advanced optimization techniques. Of these, mixed integer linear programming has become a powerful tool in which to seek an optimal solution subject to a large number of constraints and objectives. In this paper, a review of the application of MILP to water distribution networks is presented, the methodology is described, key components of the optimization process are discussed, and several case studies that illustrate how the use of MILP has successfully addressed some very practical problems are analyzed. At the end of this article, readers will have a detailed knowledge of how MILP can provide the tools to redesign or manage water distribution systems in order to reduce the costs associated with these systems as well as improve the network performance.