Developing a methodology for the design of water distribution networks using genetic algorithm

Gençoğlu, Gençer
The realization of planning, design, construction, operation and maintenance of water supply systems pictures one of the largest infrastructure projects of municipalities; water distribution networks should be designed very meticulously. Genetic algorithm is an optimization method that is based on natural evolution and is used for the optimization of water distribution networks. Genetic algorithm is comprised of operators and the operators affect the performance of the algorithm. Although these operators are related with parameters, not much attention has been given for the determination of these parameters for this specific field of water distribution networks. This study represents a novel methodology, which investigates the parameters of the algorithm for different networks. The developed computer program is applied to three networks. Two of these networks are well known examples from the literature; the third network is a pressure zone of Ankara water distribution network. It is found out that, the parameters of the algorithm are related with the network, the case to be optimized and the developed computer program. The pressure penalty constant value varied depending on the pipe costs and the network characteristics. The mutation rate is found to vary in a range of [0.0075 0.0675] for three networks. Elitism rate is determined as the minimum value for the corresponding population size. Crossover probability is found to vary in a range of [0.5 0.9]. The methodology should be applied to determine the appropriate parameter set of genetic algorithm for each optimization study. Using the method described, fairly well results are obtained.
Citation Formats
G. Gençoğlu, “Developing a methodology for the design of water distribution networks using genetic algorithm,” M.S. - Master of Science, Middle East Technical University, 2007.