The Least-Cost Design of Water Supply Networks Using the Conceptual Operator of Dynamic Threshold in GeneticAlgorithm (GA-DTO)

This paper introduces a metaheuristic optimization model for pipe networks on the basis of the notion of dynamic threshold. At first, the problem cost function, constraints and the procedure for coupling the hydraulic simulation model to the optimization are developed. Then, a simple version of binary genetic algorithm is exploited to solve the problem which is equipped with the dynamic threshold method as one of the optimization operators. By means of this operator, as the optimization progresses, the problem decision space is gradually contracted according to the search history. This increases the chance of finding the global optimum design as the search space is more and more condensed by the dynamic thresholds. The proposed scheme is then applied against two benchmark examples upon which, the method is investigated and compared with the previous studies. The results show that the incorporation of the dynamic threshold into a simple genetic algorithm can make it computationally efficient as well as more promising in finding the global optima.