Optimizing irrigation efficiency: applying NSGA-II to minimize water loss in canals

This study proposes an integrated optimization framework that combines HEC-RAS hydraulic modeling with the Non-dominated Sorting Genetic Algorithm II (NSGA-II) to enhance irrigation efficiency in the Western North Tarum Canal, Indonesia. The framework optimizes sluice gate operations under real-world constraints, including reduced capacity due to sedimentation and seasonal flow variability. Field-calibrated simulations and sensitivity analyses confirm the model's robustness. Compared to existing operations, the optimized model reduced the average annual water distribution mismatch from 0.607 m3/s to 0.018 m3/s, achieving a net reallocation of 0.137 m3/s across four irrigation segments. Notable improvements include a 0.020 m3/s deficit reduction in Pedes and a 0.022 m3/s oversupply correction in Rengasdengklok. Segment-level percentage bias (P-bias) values fell below 10%, indicating stronger alignment between supply and demand. With an average runtime of 50 minutes per optimization, the method is computationally feasible for operational planning and can be scaled to other sediment-prone. The results highlight the potential of NSGA-II-based optimization to support more equitable and sustainable water distribution in large-scale canal networks.