Various continuous harmony search algorithms for web-based hydrologic parameter optimisation

Abstract

This study compares five different harmony search algorithms that consider continuous variables inherently for the hydrologic parameter calibration. A rainfall intensity assessing model, which can provide stochastic rainfall sizes used for various structure designs, is optimally calibrated with the harmony search algorithms. The results showed that three different harmony search algorithms found better parameter values for the rainfall intensity model than mathematical optimisation technique (Powell method) and evolutionary computation technique (genetic algorithm) with respect to root mean square error (RMSE). Also, in order to enhance user-friendliness in utilising this technique, a web-based technique was developed, which optimally performs parameter calibration and presents its result on a digital map without requiring software installation.