查詢結果分析
相關文獻
- 遺傳演算法優選水庫運用規線之研究
- Determining Minimax Profile Errors of Machined Parts with Polygonal Cross-Sections
- 遺傳演算法在最佳化設計問題之應用
- 遺傳演算法於配送點選擇之應用
- Discrete Optimization of Truss Structures Using Genetic Algorithms with Adaptation
- 遺傳演算法在貨櫃船結構扭轉強度之離散最佳化設計之應用
- 以遺傳演算法作鋼桁架結構之離散最佳化設計
- 都市污水下水道管網系統規劃設計模式之建構
- 路徑基礎類粒子群最佳化演算法於求解含凹形節線成本最小成本轉運問題之研究
- Optimum Design and Manufacturing Tolerance Allocation in Product Design
頁籤選單縮合
題名 | 遺傳演算法優選水庫運用規線之研究= |
---|---|
作者 | 陳莉; 張斐章; |
期刊 | 農業工程學報 |
出版日期 | 19951200 |
卷期 | 41:4 1995.12[民84.12] |
頁次 | 頁20-30 |
分類號 | 443.96 |
語文 | chi |
關鍵詞 | 遺傳演算法; 水庫運用規線; 最佳化; Genetic algorithms; Reservoir's rule curve; Optimization; |
中文摘要 | 遺傳演算法是一種具平行演算結構的高效率搜尋方法,此法基於自然界物競天擇的理念,將搜尋點的自變數編碼成染色體,類比生物上的基因複製、交換與突變,可有效的選擇計算表現良好的點,而在每一代的繁衍中自動淘汰其他的點,達到最佳化的目標。最近則發展出實數的編碼,成功的改進了以往二位元編碼法需再解碼的不便,節省許多程式執行時間。本研究以物件導向之C++語言設計了上述兩種不同編碼表示法的運算元,並以一多峰函數為最佳化程序的測試,結果證實以實數編碼法為最佳。 本研究以淡水河流域的石門水庫為對象,檢討及修正其M-5運用規線。首先分析其運用規則及年度運轉報告,明瞭其旬運轉操作的模式;再以物件導向C++程式設計一套模擬上述操作的模式,利用36年歷史流量記錄,並以民國83年之需水情況下精確計算出下游各淨水廠與農業灌區的缺水指數;最後以遺傳演算法最佳化模式,快速而有效率的執行搜尋一組使缺水指數最少的運用規線。結果顯示遺傳演算法可有效、自主的獲致一組使目標函數達最佳的水庫操作規線。 |
英文摘要 | Based on the Darwinian natural selection, the genetic algorithm is a parallel and efficient search method. Its search space is first coded as chromosome. The basic three operators, i.e. gene reproduction, crossover and mutation, are then used to search the performance points. Each generation automatically eliminates those undesirable points for achieving the optimal goal. Recently, the real-coded method has been developed to improve the inconvenience of the binary-coded method in code convering process. This improvement saves much computer running time. The object oriented language C++ is used to implement the basic GA operators with different coded methods. A multi-peak function has been tested in the optimization processes. The results show that the real-coded method is better than the binary-coded method. The M-5 rule curve, which was constructed in several decades age, of Shi-Man Reservoir in Tan-Shui river basin is intended to be revised in this study. The operating rules and the annual operating reports are first analyzed for learning its habitual operating method precisely. The object oriented C++ language is then used to construct the simulation model. Water releasing is guided by the rule curve in this simulation model and the shortage index of the 1994 target year is calculated based on 36 years inflow record. The genetic optimization algorithm is applied for searching the rule curve with minimum shortage index. The results show that the GA method is capable of efficient and automatic obtaining the rule curve where the optimal objective function can be reached. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。