Evaluating the Efficiency of a Multi-core Aware Multi-objective Optimization Tool for Calibrating the SWAT Model

The efficiency of calibrating spatially distributed hydrologic models is a major concern in the application of these models to understand and manage natural and human activities that affect watershed systems. In this study, we developed a multi-core aware multi-objective evolutionary optimization tool, MAMEO, to calibrate the Soil and Water Assessment Tool (SWAT) model. The efficiency of MAMEO and that obtained with the sequential method were evaluated with data from the Little River Experimental Watershed. By using a 16-core machine, test results showed that calibrating SWAT with the MAMEO method required 80% less time than needed by the sequential method. MAMEO can provide multiple non-dominated parameter solutions in an efficient manner and enable modelers to simultaneously address multiple optimization objectives.