Please use this identifier to cite or link to this item:
標題: 網路規劃應用於多塔吊計劃之最佳化
作者: 何岳峰
Ho, Yueh Feng
關鍵字: Tower Crane Planning
Network Programming
出版社: 土木工程學系
摘要: 高樓建築施作,塔吊作業常是施工要徑作業,故審慎的塔吊計畫攸關工程進度延誤與否。而在塔吊配置問題中,具有解答組合空間龐大的困難,故求解優良塔吊計畫不易。全面回顧前人的研究,歸納過去主要以1)窮舉法、2)圖解法或是3)基因演算法求解,然而其方式皆有部份缺點有待改進。 本研究開發完成一套塔吊最佳化模式稱為Tower-crane Optimization Model(以下簡稱 TOM)。TOM乃應用網路規劃演算法(Network Programming),考量塔吊完成吊送時間最短為求解目標(基因演算法中則稱為:適存值)。在同時滿足供應點之供應限制量及需求點之需求滿足量的條件下,搜尋單一塔吊甚至多組塔吊之配置地點。TOM可求解供應點之最佳供應量,以達成塔吊完成任務需時最點之最佳解,進而符合施工中快速吊送物料的目標。與其他研究比較,本研究成果TOM之優點在於1)求解方式為解析解、2)求解效能快速、3)可延伸應用於多塔吊配置計劃及4)將供應點與需求點之供需量最佳化。
In high-rise building construction, the operation of tower crane has been the critical activity, and quite often causes schedule delay. The tower crane planning involves a large solution space among which it is difficult to search for an optimal solution. This study thoroughly reviewed the previous researches which could be categorized into three groups including 1) exhausted search, 2) graphical judgment, and 3) genetic algorithm. However the previous researches have drawbacks. This research proposes a new approach called Tower-crane Optimization Model (TOM) to assist the planning of tower crane. The TOM decides the optimal location to install tower crane(s), and calculates the amount of materials the demand spots should be fed by the supply spots. The TOM searches the feasible minimal time of operation, which is usually called fitness value, to achieve the optimal design of tower crane planning. The Network Programming Algorithm is adopted. The upper bond capacities of supply spots and the lower bond minimal requirements of demand spots are the constrain functions. To sum up, the TOM outperforms the previous researches in 4 respects, which are 1) analytical solution, 2) high efficiency in term of computing time, 3) extension to multiple sets of tower cranes, and 4) providing detailed delivery plan between demand spots and supply spots.
Appears in Collections:土木工程學系所



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