Optimization of Flood Control Systems by Combining Discrete Differential Dynamic Programming and Genetic Algorithm

The present study is aimed to optimum design the one of the structural methods of flood control (Levees) based on dynamic programming (DP). In order to accelerate computing time, a model was created based on the combination of discrete differential dynamic programming and genetic algorithm (GA-DDDP). The objective function of this model is to minimize the total costs, including the cost of constructing the structure and residual damage cost. For this purpose, the length of the studied river was divided into three reaches and the optimal dimensions of the four sections were obtained as output results. Modeling and running was done by coding in MATLAB software and the results showed that the running time of GA-DDDP combined model is just 11% of the time to running the classic dynamic planning model. Also in this study, the results of the four-section model were compared with the simpler single-section model (equivalent section). The results show that the simplification of model to one section causes the increasing costs to about 20%.