Particle Swarm Optimization is a well established optimization technique. Nevertheless, one of its main drawbacks comes from the fact that it is difficult to maintain acceptable levels of population diversity and to balance local and global searches. In this paper, we describe a discrete variant of PSO with increased diversity whose performance is initially investigated by applying it to a discrete, real-world problem: the design of Water Distribution Systems. Two traditional benchmark problems in the Hydraulic Engineering literature are considered: the Hanoi new water distribution network and the New York Tunnel water supply system. The obtained results exhibit considerable improvements regarding both convergence characteristics and the quality of the final solutions. A really important conclusion is that a small representative sample of the algorithm�s runs can be used to consistently achieve near optimal results at a much reduced computational cost. This is of paramount importance from an engineering perspective. Finally, to show the scalability of the model, we have applied the algorithm to a real-world water distribution network.

water distribution systems, optimal design, evolutionary algorithms, particle swarm optimization, population diversity.