遗传扩展蚁群算法用于马斯京根模型参数估计  

Genetic extended ant colony algorithm for parameter estimation of Muskingum routing model

作　　者：赵红杰[1] 柏继云[1] 马力[2] 

机构地区：[1]东北农业大学理学院,哈尔滨150030 [2]东北农业大学工程学院,哈尔滨150030

出　　处：《东北农业大学学报》2014年第8期118-123,共6页Journal of Northeast Agricultural University

基　　金：黑龙江省青年科学基金(QC2011C045)

摘　　要：文章针对扩展蚁群算法收敛速度慢,易陷入局部最优缺点,对扩展蚁群算法提出改进策略,引入遗传算法产生初始解,加入局部细搜策略。根据解的权重改进解存储器中每个解权值,增加解的方向性,快速获得最优解,通过多个典型函数寻优确定方法有效性。利用改进后算法解决洪水演算马斯京根模型参数估计问题,通过与现有马斯京根模型参数估计方法对比,验证算法具有更好优化性能,为精确估计马斯京根模型参数提供更有效方法。According to extended ant colony algorithm converging slowly and easily falling into local optimum, it presented some improved strategies：introduced genetic algorithm to produce the initial solution and join the local fine search strategy to avoid ants in local optimum and the weight of each solution improved by its＇ importance of the memory to get the optimal solution quickly and increase the direction. This paper used the improved algorithm to solve flood routing problem by parameter estimation of Muskingum routing model,by comparison with the existing parameter estimation of Muskingum routing method, validated algorithm has better optimize performance, and provide a more effective way to accurately estimating the parameters of Muskingum routing model.

关 键 词：遗传算法 扩展蚁群算法 连续空间优化 马斯京根模型 参数估计 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]