基于枝解法退火遗传算法的树状管网优化设计  

A design model of the optimal tree-type pipe network based on the branch solution-annealing genetic algorithm

作　　者：许文斌[1] 王圃[1] 何英[2] 王颖[1] 

机构地区：[1]重庆大学城市建设与环境工程学院,重庆400045 [2]长安大学环境科学与工程学院,西安710049

出　　处：《安全与环境学报》2013年第3期97-101,共5页Journal of Safety and Environment

基　　金：建设部科技项目(2001-45)

摘　　要：以管网年费用折算值为目标函数,以管径为优化参数,建立了树状供水管网的优化设计模型。根据树状管网流量的单一性,结合工程实际将支干管分明的大型树状管网按层次进行分解,再逐层采用模拟退火遗传算法对这一模型进行求解。这样可以避免因整个管网系统管段数过多、染色体过长而导致进化容易陷入局部最优解或难以收敛等缺点。同时,在算法运用过程中,对退火遗传算法进行了适当改进,使其更有效地应用到实际工程中去。分别采用传统界限流量表法、管网系统整体优化法和基于枝解法退火遗传算法的树状管网优化法对某农村供水管网进行了设计。结果表明,三者年费用折算值分别为235.21万元、217.87万元和214.89万元,相对前二者,第3种方法分别节省了20.32万元和2.98万元。因此,在支干管分明的大型树状管网设计中,枝解法优化算法可以获得更好的优化效果。The present paper intends to introduce an optimized model proposed by its authors for a tree pipe network by taking the discounted value of the annual cost as the objective function and the pipe d-i ameter as the optimal parameters. Since there is only one flow distr-i bution result of the tree pipe network, the tree pipe network model could be solved by a simulated annealing genetic algorithm when the network was decomposed into some small chunks layer-by-layer togetherwith our engineering experience. The suggested method can get rid of the local optimal solution or the difficulty of convergence evolution so as to relieve the whole network system section number of beingtoo long and too chromosome. When put into practical engineering, a great deal of improvements can be made by usingthe annealing genetic algorithm. For example, the flowing rate constraint can help to make some special processing narrower in their researching range. In addition, it is also possible for the reverse number to be used for encoding sequence to ensure that the preparation diameter be always smaller than or equal to that of the upstream one. Besides, the boundary diameters should also be encoded to facilitate the computer program control by comparing the coding values straightforwardly. In doing so, it helps to eliminate the need for decoding each pipebranch before judging the pipe diameters concerned in the range of the boundary diameters. Furthermore, we have also designed a rural water supply network by using the traditional boundary flow method togetherwith the system optimization and the branch solution-annea-l ing genetic algorithm. The application results of the data concerned prove that the discounted value of the annual cost of the three methodswere 2.352 1, 2.178 7, 2.148 9 million RMB, respectively. As compared with the first two methods, the third one can help to save 20.30, 2.98 thousand Yuan, RMB. Thus, our studies demonstrate that the branch solution-annealing genetic algorithm enjoys a great advantage in bringing about m

关 键 词：市政工程 树状管网 优化设计 层次分解 模拟退火遗传算法 

分 类 号：S274.2[农业科学—农业水土工程]