Comparison of Five Evolutionary Algorithms for Optimization of Water Distribution Networks

Abstract

In this paper, five models based on evolutionary algorithms (EAs) are introduced and compared for the optimization of the design and rehabilitation of water distribution networks. These EAs include the genetic algorithm (GA), the particle swarm optimization (PSO), the ant colony optimization (ACO), the memetic algorithm (MA), and the modified shuffled frog leaping algorithm (SFLA). A brief description of each algorithm is introduced to explain its application. A methodology is applied for the rigorous comparison of the models in terms of the optimum solution obtained, the number of objective function evaluations corresponding to the optimum solution, the effect of starting seeds on the optimum solution, and the quality of the results. A statistical analysis is carried out and then an efficiency-rate metric is determined to assess the performance of each model. The five EAs are applied to two popular benchmark networks, the two-loop network and the New York tunnels. In addition, the models are applied to a real water distribution network of El-Mostakbal City, Egypt. The results show that the PSO outperformed the other evolutionary algorithms in terms of the efficiency-rate metric and the rapid convergence to the best solution.