摘要
在马斯京根模型参数估计问题中 ,引进了绝对残差绝对值和最小准则及相对残差绝对值和最小准则 ,直接优选参数 .探讨了遗传算法的原理及本身参数优化特性 ,总结了参数的简便设置技术 ,并针对该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 ,及可能会出现过早收敛的问题 ,提出了自适应加速遗传算法 .通过实例应用 ,进行了本文方法与传统优化方法的比较 .结果表明 ,本文方法具有直观、简便、快速、适用性强等特点 。
The minimum criterion and absolute value of absolute errors and the minimum criterion and absolute value of relative errors are introduced into the direct optimum selection of parameters in the Muskingum Routing Model. The theory of genetic algorithm and the optimal characteristics of parameters are discussed; the convenient configuration technique for the selection of parameters is summarized. In consideration of the facts that the characteristics of the optimal result apparently depends on the size of the initial variation region of parameters and too early convergence may appear, an accelerated adaptive genetic algorithm is proposed. Through application in practice, the present method is compared with other conventional methods,and the result shows that the method is objective, convenient, fast and highly adaptive ,and can be widely applied to various kinds of optimization.
出处
《河海大学学报（自然科学版）》
CAS
CSCD
北大核心
2001年第4期9-12,共4页
Journal of Hohai University(Natural Sciences)
基金
教育部优秀骨干教师基金资助项目 (教技司 [2 0 0 0 ] 65- 14 )
关键词
适应遗传算法
马斯京根模型
河道洪水
演算
adaptive genetic algorithm
superior individuals
Muskingum routing model
parameter estimation
