An operative approach to water distribution system rehabilitation

The selection of the most effective course of action for pipe rehabilitation is crucial to avoid the uncontrolled deterioration of water distribution systems (WDS). Moreover, it provides a rare opportunity to modify some elements of the existing infrastructure in order to meet the present service requirements. For instance, a rehabilitation plan could be aimed at simultaneously improving system reliability, pressure levels and/or leakage reduction. In this work WDS rehabilitation is formulated as a multi-objective optimization problem where decision variables for each pipe are "do-nothing" or "replace" by selecting from a set of commercial pipe diameters. It is shown that the solutions returned by a multi-objective genetic algorithm (MOGA) in such a complex decision scenario although good, could not be easily used to support decision making. A prioritization strategy is developed and presented which is based on MOGA solutions and provides an effective tool for both analysts and decision makers.