基于退火遗传算法的给水管网优化  

Optimization of Urban Water Network by Annealing Genetic Algorithm

作　　者：王圃[1] 衡洪飞 岳健 

机构地区：[1]重庆大学三峡库区生态环境教育部重点实验室 [2]四川宜宾清源水务有限公司,四川宜宾644002

出　　处：《中国给水排水》2007年第1期60-63,共4页China Water & Wastewater

基　　金：建设部科技项目(2001-45);重庆市重点科技攻关项目(CSTC;2006AB7020)

摘　　要：给水管网优化设计的难点在于如何处理约束条件,采用较多的是遗传算法中的罚函数法,但其罚系数的选择较繁琐。故提出采用不可行度的概念来处理约束条件,然后应用退火遗传算法进行优化计算。为了缩短算法的计算时间,可在编程上进行一定的加速处理。与其他优化方法相比,该算法具有收敛速度快、精度高的优点,并在四川某市的给水管网扩建工程中得到了成功应用,取得了良好的经济效益和社会效益。Genetic algorithm is usually used to optimize the water network. But the difficulty is how to deal with constraints. Penalty function method is mainly used for optimization of constraint functions. But it is difficult to design suitable penalty factor to optimization. This study uses successfully Infeasibility Degree （ID） to deal with constraints and avoid the difficulty when using the adaptive simulated annealing genetic algorithm for optimization. The optimal algorithm is very good on astringency velocity and solution quality. And it has been used successfully in the project of urban water supply network in a southwest city of China.

关 键 词：给水管网优化 不可行度 退火遗传算法 

分 类 号：TU991.33[建筑科学—市政工程]