论文部分内容阅读
提出一种新的基于连续不确定区域的遗传算法求解水文地质逆问题。该方法运用概率与统计理论 ,将参数求解空间分为连续的不确定区域 ,并分别在各个区域利用改进的遗传算法计算各组参数对应的目标函数 ,然后通过比较目标函数在容许值内的频率分布来缩小参数的识别空间 ,从而使参数逐步达到最优。与传统的试估-校正方法类似 ,该方法实质上也是一种逐步逼近的迭代过程 ,但通过概率论方法处理 ,可大大节省传统的简单遗传算法识别参数所需的时间。实例求解结果表明该方法在参数识别过程中具有较好的效果
A new genetic algorithm based on continuous and uncertain regions is proposed to solve the hydrogeological inversion problem. The method uses probability and statistical theory to divide the parameter space into continuous uncertain regions, and uses the improved genetic algorithm in each region to calculate the objective function corresponding to each group of parameters. Then by comparing the frequency of the objective function within the allowable value Distribution to reduce the identification of space parameters, so that the parameters gradually achieve optimal. Similar to the traditional trial-and-error correction method, this method is essentially an iterative process of gradual approximation. However, by using the method of probability theory, the time required for the traditional simple genetic algorithm to identify parameters can be greatly saved. The result of the example shows that this method has a good effect in the parameter identification process