Multi-objective optimization of urban stormwater systems

Under the dual impact of rapid urbanization and climate change, flooding has become one of the major natural disasters in urban areas. Urban stormwater systems face challenges such as limited capacity, inefficient connectivity, and inadequate control measures. Traditional stormwater management systems often fail to balance flood risk reduction with cost-effectiveness. A multi-objective optimization model is developed using the Non-dominated Sorting Genetic Algorithm II (NSGA-II) to optimize stormwater systems under excessive runoff conditions. The method simultaneously addresses flood risk mitigation, facility efficiency, and cost-effectiveness across different rainfall recurrence intervals. To evaluate the effectiveness of the optimized stormwater system, a comprehensive evaluation index system is established, incorporating flood risk prevention, facility utilization efficiency, and economic costs. A case study with a 30-year return period demonstrates that increasing stormwater storage to 27 % of the cumulative upstream flood volume, optimizing inlet timing to 44 mm/h, and enhancing discharge capacity to 0.8 m³/s significantly improve system performance. The findings highlight the importance of scientifically managing key decision variables to optimize stormwater systems, providing valuable decision support for cities facing increasing flood risks.