Evaluation of direct search and genetic algorithms in optimization of muskingum nonlinear model parameters - a flooding of Karoun river, Iran

Mohammadi Ghaleni, Mehdi; Ebrahimi, Kioumars

doi:10.22059/jwim.2013.30336

Evaluation of direct search and genetic algorithms in optimization of muskingum nonlinear model parameters - a flooding of Karoun river, Iran

Authors

10.22059/jwim.2013.30336

Abstract

In this research, performance of Pattern Search algorithm (PS) was investigated in order to optimize the nonlinear Muskingum model parameters in contrast with the general algorithms, for Wilson and Karoun floods routing. Wilson flood routing has been chosen to compare the ability of PS algorithm with other algorithms and, Karoun flood routing has been chosen for a better evaluation of PS. The results of the Wilson flood routing using PS algorithm showed that the absolute and the sum squares errors were equal to 62.65 and 29.48 m3/s, respectively. Also, the lowest difference between the real and routing Wilson flood discharges, by the PS algorithm, was found to be equal to 0.29 m3/s. The results of Karoun flood routing using PS algorithm showed that the sum square error, sum absolute error and difference between the real and routing floods were equal to 7842.10, 420 and 9.7 m3/s, respectively. A comparison between the current studies with the other research results showed that PS algorithm is efficient and could be suggested for optimizing of nonlinear Muskingum model parameters.

Keywords