Implementation of a Prototype Genetic Algorithm for Optimal, Coordinated Operation of Stormwater Pump Stations

As heavy rainfall events increase, the need for efficient operation of urban drainage systems is becoming critical. In this study, we implement a prototype model designed to generate an optimal, coordinated operating schedule for an urban drainage system with the objective of minimizing system-wide flood damage. The proposed model rapidly estimates the total damage reduction by combining precomputed flood mitigation performance curves for each station with period-specific initial inundation indicators, and then selects the schedule that maximizes total damage reduction. The model is implemented as a Genetic Algorithm (GA) in a Python/DEAP environment and validated using a hypothetical pilot area. Validation using a hypothetical case confirmed that by rationally excluding inferior combinations of solutions, the optimization mechanism can be appropriately embedded in pump station operational decision-making, and its short execution time indicates the potential for deployment as a tool for real-time operation.