新型層級模糊認知圖之建構與應用──以未來汽車之技術發展為例

Construction and Application of a Novel Hierarchical Fuzzy Cognitive Map: A Case Study of Future Automotive Technology Development

模糊認知圖已廣泛應用於各類實務問題，並藉由模糊推理機制有效模擬系統之複雜動態行為，以直接影響關係為基礎所建構之鄰接矩陣，在其中扮演著舉足輕重的角色。由於鄰接矩陣並未考量因素間之間接影響關係，且多準則決策分析多以層級架構呈現問題的內涵，本研究因此於初始鄰接矩陣導入模糊決策實驗室法所決定之因素間直接與間接影響關係，並以非線性赫布學習進行優化，強化推理效果。在實務應用上，鑑於自動駕駛等技術發展在推動智慧交通的關鍵性，本研究以國內未來汽車技術發展趨勢為主軸，就應用、資源與障礙三個面向分別建立相對應之模糊認知圖，並透過關鍵因素與因果圖分析提出相關發展策略與行動方案。結果顯示，為有效提升汽車業於技術發展之績效，應優先著手改善自動駕駛技術Level 3自動化程度、電池能源儲能、市區公車全面電動化等在三個面向中的關鍵項目。本研究使模糊認知圖在層級決策分析的發展上更形完整，實證結果可支援國內汽車產業在未來汽車技術發展上制定相關決策。

The fuzzy cognitive map (FCM) has been widely applied to various practical issues and can effectively simulate the complex dynamic behavior of systems through fuzzy inference mechanisms. The adjacency matrix constructed based on direct influence relationships plays a crucial role in this process. Since the adjacency matrix overlooks indirect influence relationships among factors, and multi-criteria decision-making analysis mostly presents the problem structure in a hierarchical form, this study therefore incorporates the direct and indirect influence relationships among factors determined by the fuzzy Decision-Making Trial and Evaluation Laboratory method into the initial adjacency matrix and applies nonlinear Hebbian learning for optimization to enhance the inference performance. In practical applications, given the critical role of technological developments such as autonomous driving in advancing intelligent transportation systems, this study focuses on future trends in domestic automotive technology development and establishes corresponding fuzzy cognitive maps across three dimensions: applications, resources, and barriers. Relevant development strategies and action plans are then proposed through key factor and causal diagram analysis. The results show that, to effectively improve the performance of technological development in the automotive industry, priority should be given to advancing Level 3 autonomous driving technology, battery energy storage, and the complete electrification of urban buses, which are key across the three aspects. This study provides a more comprehensive framework for the development of fuzzy cognitive maps in hierarchical decision-making analysis, and the empirical results can support the domestic automotive industry in formulating relevant decisions regarding future automotive technology development.