A Pareto-based memetic algorithm for optimization of looped water distribution systems

Looped water distribution networks have traditionally been used in urban and industrial water supply. Nowadays, they are also being introduced in certain irrigation water distribution systems, such as in greenhouse horticultural systems. The design of looped networks is a much more complex problem than the design of branched ones, but their greater reliability can compensate for the increase in cost. Most articles found in the literature try to minimize the network investment cost, while other designing objectives are considered as constraints. This article introduces a multi-objective memetic algorithm that simultaneously optimizes the total investment cost, and also the reliability of the network in terms of total surplus power at the demand nodes. This memetic algorithm uses the Pareto-dominance concept to determine the quality of the solutions. The results obtained in two small water supply networks, and a large irrigation water supply network denote the good performance of the memetic algorithm here proposed in comparison with other well known meta-heuristics.