根据重现期P-降雨历时t-暴雨强度i的关系表,推求暴雨强度公式参数,是一个已知关系式的非线性模型参数优化求解问题.在理论分析及前人研究的基础上,把求解无约束最优化问题的变尺度DFP算法嵌入到改进浮点编码遗传算法中,得到一种基于变尺度DFP算法和改进浮点编码遗传算法的加速混合遗传算法.该方法兼顾了改进浮点编码遗传算法和DFP算法的优点,既有较快速的收敛速度,又能求得最优化问题全局精确解.经求解暴雨强度公式参数的实例表明,该法是有效的和可行的,且求解结果优于其它方法.
According to the P- t- i data, the stormy formula parameters was changed to determinate parameters of the non - liner known equation. On the basis of theory analysis and the original methods reach, based on the DFP method and real - code genetic algorithm improvably, a novel accelerating hybrid genetic algorithm was established by setting the DFP method in real - code genetic algorithm improvably, which was taken as a genetic operator that parallels to the select, crossover and mutation operators. This method is of advantages both of DFP method and genetic algorithm. It not only has a rather high convergence speed, but also can obtain a whole exact solution of the global optimization problem. The results of its applications in of resulting the parameters of the storm intensity formula show that this method is practical, efficient and superior to else methods.