التصميم الامثل لشبكات الصرف الصحي باستخدام الخوارزمية الجينية

Optimal Design of Sewer Networks Using Genetic Algorithm

  Sewerage pipeline networks are one of the essential infrastructures in modern cities serving domestic housing, manufacturing plants, hospitals, schools and other vital utility activities by disposing unwanted wastewater and preventing contamination of the water environment. In recent years, the global population has increased significantly in parallel with commercial and industrial activities. This has led to an increase in water consumption and consequential increases in the quantity of the wastewater produced, meaning that there is a need to construct new sewer networks in various places.

    This research illustrates the application of a new hybrid Genetic Algorithm with Heuristic Programming (GA-HP) technique in order to find the optimal design for sewer networks. The objective was to minimize the construction cost function, which is represented by excavation depth and pipe diameter. The proposed GA-HP model has fulfilled the optimum design task into two stages. Firstly, the Genetic Algorithm (GA) was applied to obtain the diameters needed for the preliminary design of the network. Secondly, Heuristic Programming (HP) preliminary designs were used to obtain the optimal slope for those diameters and to determine other characteristics such as the velocity, relative depth of water, excavation depths and total cost of the network.

    A MATLAB code was used to perform the GA-HP optimization model. The performance of eight different selection methods (RWS, RRWS, LRS, ERS, TRS, SUS, TOS and RMS), seven different crossover methods (One-point, N-point, Uniform, Flat Arithmetic, Intermediate and Shuffle), and different population sizes (50,100, 200, 300 and 400), have been examined using the proposed model to determine their impact on convergence behaviour. Tournament Selection method (TOS) and the One-point Crossover method proved to be the most efficient in relation to the optimal design. The proposed GA-HP model is tested using some benchmark problems of sewer networks from the literature. The results show that the GA-HP model is superior to all previous methods.

     In order to ensure the efficiency of the proposed GA-HP model for the design of large networks, it was examined with two case studies located in Karbala Holy city, and compared the cost of the manual designs with the designs obtained from the present model for networks. The saving percentages were (28.1%) and (28.45%) for relatively small and large networks, respectively.