查詢結果分析
來源資料
相關文獻
- 多目標最佳化進化法則於輕軌車懸吊系統參數及線性馬達控制增益設計之研究
- Mixed-Integer Hybrid Differential Evolution for Synthesis of Chemical Processes
- 自動導引車之模糊循跡控制系統設計與製作
- Manufacturing Cell Formation Using Genetic Algorithm VS. Neural Networks
- 自動導引車之韌性循跡運動控制器設計與製作
- A Function Approximation Neural Network with Genetic Training Algorithms
- 需量交易系統中交易訊息傳播與競價機制之設計與製作
- 臺電輸電系統裝設並聯電抗器補償最佳單組容量之選定
- 應用基因法則於電力系統最佳並聯電抗器規劃
- 基因式橢圓形體學習法之TSK模糊建模
頁籤選單縮合
題 名 | 多目標最佳化進化法則於輕軌車懸吊系統參數及線性馬達控制增益設計之研究=Evolution Algorithm and Its Application on Multi-objective Optimizations |
---|---|
作 者 | 張欽宗; 謝念中; | 書刊名 | 新新科技年刊 |
卷 期 | 1 2005.01[民94.01] |
頁 次 | 頁37-50 |
分類號 | 448.942 |
關鍵詞 | 基因法則; 進化法則; Genetic algorithms; Evolution algorithms; |
語 文 | 中文(Chinese) |
中文摘要 | 由於人工智慧(Artificial Intelligence)理論的快速發展,再加上電腦速度的增加,許多靠傳統最佳化方法較難解決的問題,目前均可迎刃而解,例如本文所探討的多目標最佳化問題,已可利用基因法則(Genetic Algorithms)及進化法則(Evolution Algorithms)解決,在實際工程上遇到的多目標最佳化問題,例如飛彈多目標追蹤導引問題、印刷電路板被動元件植入順序問題、輕軌電車懸吊系統參數設計問題及線性馬達PID+H∞/H₂控制增益設計問題等。 本文首先詳述進化法則的理論及方法,並說明其與基因法則之差異處,最後經由模擬分析,以輕軌車懸吊系統參數設計及線性馬達PID+H∞/H₂控制增益設計兩個問題,實際說明進化法則於解決具拘束限制之多目標最佳化問題的功效,希望該方法及觀念的提出能對院內有關多目標最佳化問題的解決提供一新的思考方向。 |
英文摘要 | Due to the increasing processing capability of the personal computer and, most importantly, the recent development on artificial intelligence, many traditionally unsolvable optimizations become doable. Some good examples of this kind are multi-objective tracking and guidance of missiles, mounting sequence of passive elements on the PCB, designing suspensions on a light rail vehicle, and designing PID+/ control gains of linear motors as well. Recently, evolution algorithms and genetic algorithms have been two main streams of solving multi-objective optimization problems. In the paper, the theories and methods of evolution algorithms are introduced first. Next, the differences between genetic algorithms and evolution algorithms are discussed. Finally, two cases, designing suspensions on a light rail vehicle and designing PID+/ control gains of a linear motor, are presented using computer simulation to verified the efficiency of evolution algorithms. Hopefully, this work provides a new approach toward any problems of multi-objective optimizations. |
本系統中英文摘要資訊取自各篇刊載內容。