Application of a Genetic Algorithm to Storm Sewer Network Optimization

In this paper, a genetic algorithm is developed for the optimal design of storm water networks. The nodal elevations of the sewer network are taken as the decision variables. A steady state simulation code is used to analyze the trial solutions provided by the GA optimizer. The performance of the four selection schemes namely, conventional roulette wheel, roulette wheel selection with linear scaling, roulette wheel selection with ranking and, finally, roulette wheel selection with power law scaling, is studied by applying the model to some benchmark examples in the literature. The conventional roulette wheel selection scheme produced superior results compared to other methods. The results produced by the proposed model are either comparable or superior to existing results in the literature.