Optimal parameter estimation for Muskingum model based on Gray-encoded accelerating genetic algorithm

In order to reduce the computational amount and improve the computational precision for parameter optimization of Muskingum model, a new algorithm, Gray-encoded accelerating genetic algorithm (GAGA) is proposed. With the shrinking of searching range, the method gradually directs to an optimal result with the excellent individuals obtained by Gray genetic algorithm (GGA). The global convergence is analyzed for the new genetic algorithm. Its efficiency is verified by application of Muskingum model. Compared with the nonlinear programming methods, least residual square method and the test method, GAGA has higher precision. And compared with GGA and BGA (binary-encoded genetic algorithm), GAGA has rapider convergent speed.