Application of GA-HP model for the optimal design of large sewer networks

Abstract

This research paper describes the application of a GA-HP model to the design of large gravity sewer networks. The main problem associated with the optimum design of sewer networks is achieving applicable designs that offer the lowest possible construction costs. The GA-HP model is a new technique that combines a Genetic Algorithm (GA) with Heuristic Programming (HP). One of the reasons for adopting this model is that determining the optimal design of sewer networks requires two stages: the GA can be used to obtain the optimal diameter, and then HP can be applied to obtain the optimal slope and other hydraulic characteristics. The GA was run using the tournament selection method with one-point crossover and a population size of 200. To ensure the efficiency of the GA-HP model for the design of large networks, the model was examined in two case studies, these being the medium and large networks located in the holy city of Karbala in Iraq, which contains 90 pipes and 91 manholes, and 354 pipes and 355 manholes, respectively. The cost for the applied manual designs was then compared with that for the designs obtained from this model for these networks, which indicated potential savings of 23.7% and 26.6% for the medium and large networks, respectively.