Optimizing Water Distribution in a Grid-Based Irrigation System Using Evolutionary Methods

This paper investigates the optimization of an irrigation system distributed over an agricultural area discretized into unit cells, using evolutionary algorithms for the control of water irrigation points (taps). The model simulates the distribution of water through strategically placed irrigation points, considering the individual requirements of each cell. The main objective is to minimize the difference between the amount of water needed and delivered, while reducing the total consumption. The dynamics of fitness over generations are analyzed, as well as the average behavior of deficit, surplus, and relative humidity. The results highlight a relatively uniform distribution of delivered water and a stable convergence of the fitness function, demonstrating the efficiency of the proposed method in managing water resources in a sustainable way. In this matter, compared to the full-activation scenario, the presented model reduced total water use by more than 50%, achieving zero deficit, minimal surplus, and a 46% improvement in overall fitness. Although the approach demonstrates promising results in simulated scenarios, it does not currently incorporate real-time sensor data or field validation, which are planned for future development. The study provides a solid basis for the development of smart irrigation systems, adaptable to the variability of soil and climatic conditions.