Sequential Multi-objective Evolutionary Algorithm for a Real-world Water Distribution System Design

This paper presents a methodology based on a three-stage multi-objective optimization model for solving the problem of Battle of Background Leakage Assessment for Water Networks (BBLAWN) at WDSA2014 conference. At the first stage, the optimal design of pipeline rehabilitation, pump scheduling and tank sizing is formulated and solved on the skeletonized network by a optimizing (1) the costs of pipes, pumps and tank upgrading and (2) the cost of water losses and energy. Three optimal solutions are used for a second optimisation step on the full network (i.e. not skeletonised). The third optimisation step is then performed starting from second stage optimal solutions considering the three objectives of the original problem.