基于BOD-DO水质模型参数反演的小生境蚁群算法  

Niche ant colony algorithm for parameter inversion of BOD-DO water quality model

作　　者：张红菲[1] 张新明[1] 

机构地区：[1]哈尔滨工业大学深圳研究生院,广东深圳518055

出　　处：《水力发电学报》2013年第6期92-98,共7页Journal of Hydroelectric Engineering

基　　金：国家自然科学基金(4037404);高等学校博士学科点专项科研基金(新教师)(20102302120072);深圳市科技研发基金基础研究计划(JC201005260139A)

摘　　要：针对传统蚁群算法容易陷入局部最优(即"早熟"现象)的问题,本文将小生境方法和蚁群算法相结合,提出了基于适应值共享的小生境蚁群算法。我们首先将其应用于多模态函数优化问题,并将寻优结果与传统蚁群算法以及小生境遗传算法进行对比,验证了该算法的有效性和较高的精确度。其次,基于水质模型,将改进的小生境蚁群算法应用于耗氧系数、复氧系数和BOD沉浮系数的参数反演计算。数值模拟结果表明该方法是一种精度高、收敛速度快且易于计算机实现的有效方法。In this article, a parameter inversion problem of water quality model is considered. Combining the ant colony algorithm with fitness sharing, we presents a new niche ant colony algorithm based on the fitness sharing principle to overcome the ＂premature＂ phenomenon of traditonal algorithm. First, we applied this new algorithm to multi-modal function optimization and compared the results with those calculated by the traditional algorithm and a niche genetic algorithm. The comparison shows that the new method is more effective and more accurate in global optimization. Then, we estimated the parameters （oxygen consumption coefficient, reaeration coefficient and BOD checkered coefficient） using a BOD-DO water quality model. The results show that the method is more accurate, faster in convergence and easy to implement numerically.

关 键 词：水质模型 参数反演 小生境蚁群算法 适应值共享 

分 类 号：X522[环境科学与工程—环境工程]