单一重现期非线性暴雨强度模型参数的确定——新型混合加速遗传算法的构建与应用

Parametric determination of nonlinear storm intensity model with single return period:construction and application of a new hybrid accelerating genetic algorithm

对浮点编码遗传算法进行了改进,并嵌入Powell方向加速局部搜索算法与加速循环操作,从而构建了新型混合加速遗传算法。确定单一重现期的非线性暴雨强度模型参数的应用实例表明,该方法兼顾了改进浮点编码遗传算法和Powell算法的优点,是一种既可以较大概率、快速地搜索全局近似最优解,又能进行局部细微搜索的出色的非线性优化方法。该法在其它工程问题的非线性模型参数的确定中也具有广阔的应用前景。

Based on improvement of real-code genetic algorithm and insertion of Powell′s local search algorithm and accelerating cyclic operation, a novel accelerating hybrid genetic algorithm(NHAGA) was established. NHAGA was used to determining the parameters of the nonlinear storm intensity model with single return period and the results show that this method,having both the advantages of Powell method and genetic algorithm,is a excellent nonlinear optimization method,which not only may obtain a integrate exact solution of the global optimization problem with a rather high convergence speed, but also can partly carry out the fine search that this method is practical, efficient and superior to others. It has the wide application prospect in determining parameters of the nonlinear models which has been encountered in else project.