Optimum designs of drinking water networks

V SUMMARY The aim of this study is to investigate the conformity of dead-end method, a project method of drinking-water networks, with genetic algorithm method, an optimisation method which gets to be developed in the last decades. Network calculations depend on very complex methods, because of this it is estimated that flexible optimisation methods, like genetic algorithm, will show more conformity with these methods as well. At this point; natural parameter set of the problem (diameters of pipes; genes) has been prepared, the limits (minimum and maximum values according to the dead-end method) have been determined, the reproduction operator has been formulated and, as a conclusion, the applied project results and the sample results have been compared. Reproduction operator has been formulated as >reproducöon=(fi / lf]Y- Here ; fi : multiplication of the length and the diameter of the nominally longest pipe, Ifj : sum of the multiplication of the length and the diameters of all pipes in the network, ^reproduction: the ratio of fi to IS \ ? This ratio inclines to get bigger as the diameters get smaller. So, maximization of this value (within limits) will bring the optimum result as well. The first variable (pipe diameters; string of genes) that is randomly determined gets to be mutated step by step, new combinations (chromosomes; string of genes) are obtained and by taking the greatest ratio of reproduction operator as basis, the pressures are balanced. But, it must be noted that, the resistance of the connection parts has been neglected in the sample solution. This study just includes the investigation of the conformity of dead-end method and genetic algorithm as a first step. As a result of the study, it has been defined that there is a conformity between the dead end method and genetic algorithm optimization method. Key Words : Dead-End Method, Genetic Algorithm.