Search Space Reduction for Genetic Algorithms Applied to Drainage Network Optimization Problems

In recent years, a significant increase in the number of extreme rains around the world has been observed, which has caused an overpressure of urban drainage networks. The lack of capacity to evacuate this excess water generates the need to rehabilitate drainage systems. There are different rehabilitation methodologies that have proven their validity; one of the most used is the heuristic approach. Within this approach, the use of genetic algorithms has stood out for its robustness and effectiveness. However, the problem to be overcome by this approach is the large space of solutions that algorithms must explore, affecting their efficiency. This work presents a method of search space reduction applied to the rehabilitation of drainage networks. The method is based on reducing the initially large search space to a smaller one that contains the optimal solution. Through iterative processes, the search space is gradually reduced to define the final region. The rehabilitation methodology contemplates the optimization of networks using the joint work of the installation of storm tanks, replacement of pipes, and implementation of hydraulic control elements. The optimization model presented uses a pseudo genetic algorithm connected to the SWMM model through a toolkit. Optimization problems consider a large number of decision variables, and could require a huge computational effort. For this reason, this work focuses on identifying the most promising region of the search space to contain the optimal solution and to improve the efficiency of the process. Finally, this method is applied in real networks to show its validity.