Multiobjective Parameter Calibration of a Hydrological Model Using Harmony Search Algorithm

The tank model among various deterministic Rainfall-Runoff (RR) models is often preferred for its simple concepts by many previous studies. However, it requires much time and effort to obtain better results owing to the need to calibrate a number of parameters in the tank model. The demand for an automatic calibration method has been increasing. The success of heuristic optimization algorithms enables many researchers to focus on the other aspect of the various objective function for tank model rather than parameter calibration. In this study, Multi-objective Harmony Search Algorithm was performed for an automatic calibration. The proposed study enables parameter calibration of four storage types of tank model with 14 parameters and six scenarios based on the four objective functions. A proposed tank model, which determines optimal solution (e.g., parameters of tank model; calibration results) under four different objective functions, respectively, were compared to demonstrate tradeoff relationship between measurements data and observation data for a Daecheong dam watershed in South Korea. The proposed tank model with six different scenarios could be a successful alternative RR model with its increased accuracy and Pareto solution.