基于遗传模拟退火法的马斯京根方程参数估计

提出一种用遗传退火算法估计马斯京根方程参数的新方法,在遗传算法中融入模拟退火算法,在每一代群体产生后,对各个个体独立地进行模拟退火过程,以其结果再作为下一代群体的个体,避免了简单遗传算法容易提前收敛的缺陷以及模拟退火法搜索较盲目的缺点,应用实例表明方法简便,直观,可广泛应用于解决多种模型的优化问题,特别在洪水预报方面有很好的应用前景。

Approach to parameter estimation of Muskingum routing model based on genetic simulated annealing algorithm

An approach-genetic simulated algorithm is presented for optimizing the parameters of Muskingum routing model. By integrating the simulated annealing into the genetic algorithm, applying simulated annealing to every individual to the GA parent generation and generating the new generation, the method can solve certain problems, such as premature convergence in simple genetic algorithm or searching blindly in simulated annaling. It has proved to be efficient and simple in some cases. Moreover, it is more easily understood than before. It can be widely used in varieties of optimization model, especially in flood forecast calculation.