知識融合一直是知識工程中極具挑戰性的研究議題，本論文的目標是探索啟發式最佳化方法對於模糊知識融合之成效，以往有很多啟發式最佳化方法已經被提出來解決最佳化的問題，而這些方法最有趣的特性就是奠基於自然現象與生物行為。令人振奮地，野貓的天性和其獨行者般的智慧觸發本論文設計一個創新的啟發式最佳化演算法，本論文提出了貓最佳化演算法以及建立一個以貓最佳化演算法為核心的模糊知識融合系統；其中，知識編碼與知識最佳化機制將同時融合多個模糊知識庫。為了證明貓最佳化演算法的效率，本論文利用著名的非線性函式將貓最佳化演算法與其他啟發式演算法做實驗比較，實驗結果證實了其高搜尋效益；在實證研究中，本論文利用程式設計學習風格診斷知識之融合應用案例來分析驗證此模糊知識系統的系統效益與使用者接受度。Knowledge fusion is always a challenging research topic in knowledge engineering area. The objective of this dissertation is exploring heuristic optimization method to succeed fuzzy knowledge fusion. Several heuristic optimization methods had been proposed to solve optimization problems previously. The most interesting characteristic of these methods is that these methods are based on natural phenomena and animal behavior. Excitingly, the instincts of wild cats and loner intelligence trigger the dissertation to devise a novel heuristic optimization algorithm. This dissertation proposed a CatOpt algorithm and developed a CatOpt-based fuzzy knowledge fusion system. Among which, the knowledge encoding and knowledge optimization can fuse multiple fuzzy knowledge bases simultaneously. To demonstrate the effectiveness of the CatOpt algorithm, the dissertation compared this method with other heuristic methods by using well-known nonlinear equations. The experimental results confirmed the high search performance of the proposed algorithm. In the empirical study, application of fusion of programming learning style diagnosis knowledge was used to demonstrate the system efficacy and user acceptance.
