摘要
The calibration and selection of conceptual hydrological model parameters is an important but complex task in runoff forecasting. In order to solve the calibration of conceptual hydrological model parameters, a multi-objective cultural self-adaptive electromagnetism-like mechanism algorithm (MOCSEM) is proposed in this paper. The multi-objective parameter calibration method of runoff forecasting avoids the “averaging effect” and considers both large and small runoffs hydrological features. In this paper, the self-identifying parameter combination forecasting method (SPCFM), a universality combination forecast model, is developed innovatively to improve forecasting precision by using the extreme parameters of Pareto optimal solutions. Finally, MOCSEM is combined with SPCFM to calibrate the parameters of forecasting model and forecast runoff of Leaf River. The results indicate that the proposed methods improve forecast accuracy and provide an effective approach to runoff forecast.