利用NSGA-Ⅱ算法求解供水管网优化改造模型  

Optimal rehabilitation model of water supply network with non-dominated genetic algorithm-II

作　　者：金溪[1] 高金良[1] 张杰[1] 王芳 

机构地区：[1]哈尔滨工业大学市政环境工程学院,哈尔滨150090 [2]武汉市城市规划设计研究院,武汉430017

出　　处：《哈尔滨工业大学学报》2008年第12期1969-1976,共8页Journal of Harbin Institute of Technology

基　　金：黑龙江省自然科学基金资助项目(ZJG0503)

摘　　要：为提高供水管网优化改造模型的客观性,给出更合理的优化结果,对供水管网改造单目标优化模型进行适当处理,将水力约束条件转化为独立的目标函数,建立供水管网改造多目标优化模型.利用面向多目标优化问题求解的非控制排序遗传算法-II(NSGA-Ⅱ)求解多目标管网优化改造模型.通过算例验证,算例管网中低压节点问题、管段负荷过大问题、管段改造投资问题,由于都作为目标函数进行求解,给出综合考虑三方面问题的优化结果.通过多目标建模思想以及面向多目标问题优化算法(NSGA-II)的引入,解决单目标模型无法描述管网改造为多目标问题的矛盾,克服采用权重系数或惩罚函数带来的不确定因素.并通过引入人工诱导基因变异算子,加快种群向可行解域的收敛速度,提高算法的收敛速度,而且改善解的合理性.To improve the objectivity of rehabilitation model for water supply network and provide more feasible solutions, a multi-objective rehabilitation model was developed by transforming the hydraulic constraints of the single objective model into objective functions, and the non-dominated sorting genetic algorithm - Ⅱ （NS- GA - Ⅱ） was used to solve the developed model. The test on a case of water supply network shows that a solution satisfied with all objectives can be obtained by considering the low pressure node, high load pipe and rehabilitation cost as objectives of rehabilitation model. The introduction of multi-objective concept and multi-objective oriented algorithm（ NSGA - Ⅱ ） into the solving process of rehabilitation problem for water supply net- work overcomes the conflict between the rehabilitation model with one objective and that with multi-objectives, which avoids the uncertainties brought by using weight coefficients or punish functions. The introduction of artificial induction gene mutation operator accelerates the convergence speed of population, thus improves the convergence speed, which proves the feasibility of the method.

关 键 词：供水管网 优化改造 非控制排序 遗传算法 人工诱导变异 

分 类 号：TU991.33[建筑科学—市政工程]