改进遗传算法在水文模型参数优选中的应用

An improved genetic algorithm for parameter optimization of hydrology model

为改善传统水文模型参数优选算法的性能,针对遗传算法的寻优效果明显依赖于模型参数的初始变化区间的大小,并且局部搜索能力较差、可能会出现过早收敛等问题,在遗传算法中加入局部搜索算子和加速算子,并引入了生物学中的小生境概念,提出了基于小生境技术的混合加速遗传算法(HAGA).该算法在广西合浦水库流域的洪水预报中得到成功应用.结果表明:基于小生境技术的混合加速遗传算法不仅有较好的全局优化性能而且精度较高,是一种既可以较大概率搜索全局最优解,又能进行局部细致搜索的优秀非线性优化方法.

In order to improve the capability of the traditional algorithm of parameter optimization in hydrology model, a hybrid accelerating genetic algorithm （HAGA） based on niche technique is suggested by adding locally searching operator and accelerated searching operator in real coding genetic algorithm. This method could avoid the results of GA excessively depended on the initialization and convergence at a much earlier stage. The HAGA is successfully applied to the parameter optimization model for the Hepu basin in Guangxi Zhuang Autonomous Region of China. The results show that this method has not only better optimization capability, but also better accuracy. It is a superior nonlinear optimal method which could locally search the global solution for greater probability.