Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/5531
標題: 螞蟻演算法與禁忌搜尋法混合模式於配水管網設計最佳化之應用
The Applications of Ant Algorithms - Tabu Search Hybrid Model on the Optimal Design of Water Network System
作者: 劉盈利
Liu, Ying-Li
關鍵字: water distribution network;配水管網;heuristic algorithms;optimization;tabu search;ant algorithms;ant-tabu;啟發式演算法;最佳化;禁忌搜尋法;螞蟻演算法
出版社: 環境工程學系
摘要: 
由於社會變遷、人口成長、都市更新及社區的開發,人們對水資源的需求也日益增多。自來水建設工程對於經濟及都會區發展、提高民眾生活水準方面占相當重要的影響,而自來水建設工程,因都市人口集中,需水量與日俱增,工業發展需求勢必更趨迫切。自來水建設工程可分集水、輸水、抽水、淨水及配水等項目,其中以配水工程一項的建設費占大宗,故若慎重規劃設計,將能節省許多的工程費用。
配水管網最佳化設計屬大尺度、困難且複雜的問題,以往採用傳統技術方法曠日費時,容易造成研究資源的浪費,近十幾年發展的啟發式(heuristic)演算法具健全且高效率的搜尋能力,對求解複雜問題能有相當優良的表現,本研究結合螞蟻演算法(Ant Algorithms)及禁忌搜尋法(Tabu Search)成為一套具有優良搜尋能力之Ant-Tabu模式,應用於配水管網最化設計上,供配水管網設計決策之參考。
本研究求解5個配水管網最佳化設計的案例,並與其他相關研究作綜合性探討分析,研究結果顯示,兼具螞蟻演算法與禁忌搜尋法之優點的Ant-Tabu模式,在每個案例皆獲得優良的配置組合,並在Hanoi及New York兩案例突破先前有關研究所提供的最佳管徑組合方案,另隨著管網最佳化設計問題趨於複雜龐大,禁忌搜尋法雖然獲得優良的建置成本,Ant-Tabu則呈現比禁忌搜尋法更優秀的表現。

The optimizations of water distribution network designs are often lager scale and highly complicated problems. Conventional optimization techniques are easily trapped at local optima and unable to successfully solve such problems. In the past decade, heuristic algorithms which have robust and efficient global search capability have been employed to solve these problems and demonstrated well performances.
This research combines ant algorithms and tabu search (TS) to form a new hybrid model, Ant-Tabu, to optimize the designs of water distribution networks. The results show that both TS and Ant-Tabu are able to achieve least-cost solutions better than those published in literature so far. It can be concluded that Ant-Tabu is a robust algorithm and suitable for optimizing water distribution network designs and other complicated problems.
URI: http://hdl.handle.net/11455/5531
Appears in Collections:環境工程學系所

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.