改进遗传算法在高非线性水质模型参数估值中的应用研究  

Research on Application of Modified Genetic Algorithm in Parameter Estimation of Highly Non-linear Water Quality Model

作　　者：冯良记[1] 唐军[1] 

机构地区：[1]大连理工大学海岸和近海工程国家重点实验室,辽宁大连116024

出　　处：《水利与建筑工程学报》2009年第2期11-13,共3页Journal of Water Resources and Architectural Engineering

基　　金：国家重点基础研究发展计划(973计划)项目(2005CB724202)

摘　　要：在标准遗传算法的基础上,提出了一种采用精英保留策略、小生境技术、适应函数调整,同时又能自适应地改变交叉和变异概率的改进遗传算法,使得算法在高维复杂水质模型多参数估值搜索时,不丢失最优解空间和后期有效分辨最优适应度。以测试函数Rastrigin为验证,得到了已知的最优结果;最后,以下水道高非线性水质模型的参数优化估值问题为实例进行验证,将优化后的水质参数代入模型中,模拟所得结果与给定的实测值吻合良好,实现了高维复杂水质模型多参数的同时估值优化功能。该算法对其他高非线性水质模型参数优化问题同样具有较好的适用性。Based on the standard Genetic Algorithm（GA）,a modified GA is presented according to the weakness of the standard GA by using the elitist reservation,niching method,adaptive function adjustment and automatically adaptive regulation in the probability of crossover and mutation.The effectiveness of improved GA（IGA） in optimization is firstly proved by applying in Rastrigin test function.This test shows that IGA is useful in nonlinear optimization.Then,IGA is applied in optimization of a complicated water quality model in sewers,which is composed of as many as 12 equations and 23 parameters.Among these 23 parameters,some are very sensitive.By using IGA,a new series of parameters is found,simulating with which the water quality model gets very close results to the measured data.It is shown that IGA is useful and successful in optimizing the parameters of other high non-linear water quality models.

关 键 词：遗传算法 优化 水质模型 参数估值 

分 类 号：X824[环境科学与工程—环境工程]