MULTI-OBJECTIVE ALGORITHMS APPLIED TO THE REHABILITATION OF WATER DISTRIBUTION NETWORKS
Water supply; rehabilitation; evolutionary algorithms; hybrid algorithm; Artificial Immune System, Particle Swarm Optmization; Tabu Search.
High urbanization intensifies the use of available natural resources for water supply. This intensification of the demand is added the deterioration of the network itself due the age or poor quality of the materials used, causing water distribution networks (WDNs) to reach critical levels of supply. As a result, there will be an increase in losses due to leaks and failures in network components, increased maintenance and operation costs, interruptions in supply and decreased reliability. The WDN rehabilitation project continues to be the main alternative for solving these problems. The optimization problem of rehabilitation projects has multiple objectives, whose solving complexity, in line with the quality of the solution found and the processing time, become one of the great challenges to be overcome by mathematical algorithms. That said, this research aims to explore the Artificial Immune System Multiobjective (MAIS), Particle Swarm Optimization (PSO) and Tabu Search (TS) as promising new options for optimization methods, based on multiobjective evolutionary algorithms, for the rehabilitation of networks water distribution. Analyzing and comparing them through evaluation metrics. In addition to proposing a new hybrid algorithm based on the Multiobjective Artificial Immune System (MOAIS), Particle Swarm Optimization (PSO) and Tabu Search (TS) methods, seeking to improve the search process, providing effective solutions with lower computational cost. The methods will be applied in two networks present in the literature and in two other real cases of city networks in the interior of the state of Pernambuco, Brazil.