Application of Evolutionary Optimization Algorithms for Rehabilitation of Water Distribution Networks

Abstract

Deteriorated water distribution networks require significant investments to maximize their functionality. The problem is that limited financial resources are allocated for rehabilitation strategies. This deficiency highlights the importance of developing a tool that helps decision makers develop maintenance and replacement management plans. The optimization tool is employed using two evolutionary algorithms: genetic algorithms and particle swarm optimization. The efficacy of the developed model is demonstrated through its application in a case study of Shaker Al-Bahery, Egypt. Furthermore, evaluation metrics are considered to compare the performance of the aforementioned algorithms. The results reveal that the particle swarm optimization exhibited superior results when compared with the genetic algorithms. Moreover, the following two multicriteria decision-making techniques are used to provide a ranking for the near-optimum solutions: multiobjective optimization on the basis of ratio analysis and technique for order preference by similarity to ideal solution. Finally, the Spearman correlation coefficient is utilized to assess the correlation between rankings obtained from different decision-making methods. The results indicate a very strong relation among the aforementioned techniques.