演化式模糊神经推论系统於大地工程决策之应用.doc

演化式模糊類神經推論系統於大地工程決策之應用 鄭明淵 柯千禾 張文德 國立台灣科技大學營建工程系教授 國立台灣科技大學營建工程系管理組博士候選人 摘要 大地工程方面的問題具有不確定、模糊與不完整資訊的特性，因此，在解決相關問題時多仰賴該領域專家經驗與知識進行決策，本研究的主要目的為模擬人類大腦決策過程，發展一「演化式模糊類神經推論系統」，透過過去之工程案例與經驗，學習累積專家決策過程與分析邏輯，輔助大地工程專家進行決策，以提昇大地工程決策的有效性與準確性。首先，本研究融合基因演算法、模糊邏輯與類神經網路，建立「演化式模糊類神經推論模式」，模式中應用基因演算法模擬生物演化過程，經由世代演化搜尋所有模糊類神經網路所需要的最佳參數；模糊理論用於模擬人類推論以及處理問題的模糊與不確定性的機制；而類神經網路則用於效仿人類學習模式。然後，本研究運用物件導向相關技術將「演化式模糊類神經推論模式」電腦化為「演化式模糊類神經推論系統」。最後，本研究使用兩個問題驗證本研究所發展的「演化式模糊類神經推論系統」可有效輔助大地工程專家進行決策分析。 關鍵字：人工智慧、基因演算法、模糊邏輯、類神經網路、演化式模糊類神經推論模式、物件導向系統分析與設計、演化式模糊類神經推論系統、大地工程決策。 EVOLUTIONARY FUZZY NEURAL INFERENCE SYSTEM FOR DECISION-MAKING IN GEOTECNICAL ENGINEERING CHENG MIN-YUAN KO CHIEN-HO CHANG WEN-TE Department of Construction Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan ABSTRACT Problems in geotechnical engineering are full of uncertain, vague, and incomplete information. Therefore, to solve most problems in geotechnical engineering depends on human experts’ decision-makings. The primary objective of this research is to imitate decision-making mechanism of human brain to develop an Evolutionary Fuzzy Neural Inference System (EFNIS) for assisting geotechnical experts to make proper decisions. Firstly, this paper fuses Genetic Algorithms (GAs), Fuzzy Logic (FL), and Neural Networks (NNs) to build an Evolutionary Fuzzy Neural Inference Model (EFNIM). In the formulated model, GAs is primarily concerned with optimizing parameters required in fuzzy neural networks (FNNs); FL with imprecision and approximate reasoning; and NNs with learning and curve fitting. Then, this research integrates the EFNIM with Object-Oriented (OO) computer technique to develop the EFNIS. Finally, the application potential of the proposed system in assisting geotechnical decision-making is validated using two real problems. KEYWORDS：Artificial Intelligence, Genetic Algorithms, Fuzzy Logic, Artificial Neural Networks, Evolutionary Fuzzy Neural Inference Model, Object-Oriente