Simulation and Optimization Models for Flood Control by Genetic Algorithm: Application to Karkheh Dam

The paper demonstrates a simulation-optimization framework for enhancing the real-time flood control with gated spillways at places where no flood forecasting data is available. A multi-objective modeling scheme is presented for flood management in a gated spillway in which the operator may specify the priorities on floods based on their different return periods. Two different operation strategies are devised. Both of the operating strategies employ ten-stage policies, which mainly rely on the reservoir water level as the input data. The second strategy benefits from both observed reservoir water level and flood peak. The optimal values of the models parameters are obtained using genetic algorithm. This is a novel approach because none of its policies needs flood forecasting data, thus, making them adaptable to any flood with any return period. To evaluate the performances of the proposed models, the flood control through gated spillway of Karkheh reservoir is considered where flood hydrographs with different return periods are routed through the reservoir.