Decentralization-based optimization of detention reservoir systems for flood reduction in urban drainage areas

The paper introduces a decentralized approach to optimizing the flood reduction efficiency of multiple detention reservoirs in urban areas. The flood reduction effects of a detention reservoir system were regionally assessed, and a new model was then created using an optimization technique which involved coupling extraordinary particle swarm optimization with a flood simulation engine, Storm Water Management Model, EPA-SWMM, to identify optimal outlet designs for a detention reservoir system. When applied to the To Lich drainage network (Hanoi, Vietnam) the proposed method outperformed the present design, reducing flooding under the design flood event. The performance of the approach was verified by comparing it with the overall objective function of total flooded volume for different design floods. The better performance of the decentralized optimization approach described here, in terms of local and regional flood management, illustrated its effectiveness for urban flooding reduction.