Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/4944
標題: 平行遺傳演算法以電腦叢集為工具應用於地下水優選問題之探討
Implementation of a Parallel Genetic Algorithm on PC Cluster to Solve Groundwater Optimization Problems
作者: 嚴浩哲
Yen , Hao-Che
關鍵字: parallel genetic algorithm;平行遺傳演算法;PC Cluster;groundwater optimization problems;電腦叢集;地下水優選問題
出版社: 環境工程學系
摘要: 
電腦叢集為高速計算領域的新寵,其最大特點為擁有極佳的價格/效能比,而應用領域涵蓋相當廣泛包括天文、物理、流體力學、電磁、氣象、水資源規劃等,面對處理問題越來越複雜之今日,許多模擬需處理大量資料,並執行龐大的運算,以得到正確或近似之答案,這種結果往往必須具備時效性,而環境污染是具有時變性的,作為一個環境決策者應體認如何在合理的時間內得到適當決策之重要性。
基於此踏入高速運算的領域已是時勢之所趨,電腦叢集亦成為一般研究團體或個人最佳的運算利器,本研究對廣範成功應用於各領域之遺傳演算法作為求解地下水復育優選問題之工具,並進行平行化之探討,藉由結合高速運算概念,期能快速求解複雜之環工優選問題。
研究結果顯示本研究所發展之平行遺傳演算法,可有效率地求解地下水優選問題,縮短大量的求解時間,在增加處理器時優選模式可維持一定的效率與加速,因此對於大尺度、複雜性高之問題,增加處理器數目便可縮短大量求解時間並求得良好品質之解。

PC Cluster is a new technique of high speed computing and has a very good cost/efficiency ratio. It has been successfully applied in many researches such as astronomy, physics, hydrodynamics, electromagnetic, meteorology and water resources planning and management. This main objective of this research is to develop a parallel genetic algorithm that can be executed on a PC Cluster platform to solve the optimal solutions of groundwater remediation problems.
The results showed that the parallel genetic algorithm can solve the complicated groundwater optimization problems effectively and more efficiently. Compared with the sequential genetic algorithm models, the computational time of the parallel version is significantly reduced. Furthermore, when the number of CPU increases, the model can still maintain its computational efficiency and speedup at a good quality. Therefore, the computational time of those large scale and complicated problems can be remarkably decreased by using the PC Clusters which contain more CPUs.
URI: http://hdl.handle.net/11455/4944
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.