Assessing the Performance of a Routing Procedure Based on the Genetic Algorithm

The so-called hydraulic routing methods incorporate both continuity and momentum-conservation equations to describe the configuration of a flood wave through a river reach. These may be very complex numerical schemes requiring, in most cases, a substantial amount of field data. Alternatively and due to their simplicity, the so-called hydrologic routing methods have been extensively used by engineers. These, such as the Muskingum method, employ only the continuity equation and a linear relationship between river storage and inflow/outflow discharges. Recently, optimization techniques, such as the genetic algorithm, have been used in order to identify the parameters of a flow routing model, which may include a non-linear relationship between river storage and inflow/outflow discharges, though keeping the general simplicity of the hydrologic routing methods. Genetic algorithms use reproduction schemes similar to those occurring in natural genetics to find solutions to a variety of mathematical problems. In the case of flow routing, the problem is one of error minimization. This flow routing model is applied to the São Francisco river basin, located in Southeastern Brazil. The case study has been selected to reflect flow conditions which may be considered appropriate for Muskingum-like routing methods.