Optimal Designs of Flood Mitigation Measures Considering Loss of Lives Using a Bi-Objective Genetic Algorithm

Using structural and nonstructural measures for flood damage reduction is a long-standing problem in water resources planning and management. In present study, an algorithm is presented to optimal design of flood mitigation measures in the watershed scale by simulation based optimization approach. To do so, the numerical model of MIKE-11 was used to calculate the risk of death and the physical damages of flood scenarios under different combinations of structural and non-structural flood mitigation measures. The numberdeath depends on the flood characteristics (the intensity of hazard) obtained by run of numerical model in each round, the vulnerability of population and the vulnerability of properties in flood-prone areas. Each of flood mitigation measures changes one of these three factors and affects on the number of death in flood events.As likewise, each measure may affect on physical damages by altering the magnitude of floods or vulnerability of land uses.In fact for a flood scenario, the physical damages in each combined option are computed using the change of geometry of rivers in model and its boundary conditions or modifying the damage-elevation curves or both. The numerical model was coupled with the NSGA-II multi-objective optimization model to provide the optimal Pareto front solutions considering two conflict objectives of minimizing the investment costs and the potential flood damages in the watershed.Finally, the presented model was applied for a small watershed in the center of Iran as a case study and the optimal trade-off solutions were calculated for different flood scenarios. Results showed by application of presented approach the investment costs may decrease several times and at the same time potential flood damages are minimized.Using a numerical model in the structure of proposed framework provides a flexible tool to consider the interaction of different measures in different reaches of watershed. For example model can dynamically predict in which rivers there is a need to build dam and what is the optimal height of dams in various branches to prevent the synchronizing the flood peaks of branches or detention dams with which measures must be applied to minimize the total cost and flood damages. For the study area, flood wall option just proposed by flood warning option in obtained optimal designs. This shows“flood wall”option to reduce damages on properties with “flood warning” option to reduce the loss of lives is an effective flood mitigation strategy for the study area. Also, if the investment level is low, the application of two non-structural measures of “waterproofing” and “watershed management measures” have a priority than the other measures for the study area. Using the obtained trade-offs, for each level of funding, decision makers can assign the optimal combined option considering the decision criteria.