Evolutionary computation techniques for optimal response actions against water distribution networks contamination

Access to safe drinking water has always been a major concern for societies. Hence, a continuous assessment of water distribution networks is required to ensure their proper performance and detect accidental contamination incidents or intentional attacks. Contamination management and return to normal condition is a quite challenging problem considering the diversity and complexity of real water distribution systems (WDS). Contamination restriction and subsequent flushing have been proposed as the most applicable solution to protect public health and prevent irreversible social, economic, and environmental damages. By detection of the contaminant existence in a WDS, the cleaning strategies should be adopted at the earliest opportunity to eliminate the public exposure to potentially harmful substances. The stressful situation and complex nature of WDSs make it practically impossible for the utility manager to identify such strategies at the moment; hence, appropriate guidelines or analyzing engines must be provided for the decision makers, in advance. In this regard, the evolutionary algorithm-based optimization models have been extensively developed in the literature to determine applicable cleaning strategies in a form of optimal contamination response actions. Genetic algorithms (GA) can be mentioned as the most general paradigm of evolutionary algorithm (EA) being utilized in WDS contamination studies. Considering the WDS contamination as a multidimensional problem, the nondominated sorting genetic algorithm-II (NSGA-II) has been the most widely used algorithm within the developed multiobjective simulation-optimization frameworks. Studies have stated that the NSGA-II capabilities allow the managers to thoroughly explore the complex WDS response action optimization problem, and provide a variety of desirable solutions to mitigate the WDS contamination.