知識“溶化”技術及無搜索快速求解智能系統的建立

Technique of Knowledge Dissolved in Neural Network and Non-searching Fast Inference

摘要: 介紹了一個模糊神經網絡專家系統.研究了知識的“溶化”技術，即導入算法、無搜索快速推理、尋出算法，井對用三分法確立輸入變量隸屬函數及模糊規則求精作了有益的研究.用以上理論對連續推鋼式加熱爐進行控制，控制效果較為理想.

Abstract: A method of building expert systems with fuzzy-neural technique is proposed.The technique of the knowledge dissolved in a neural network non-searching inference have been researched. The membership functions for the input virable have been generated by a triple separation. The simulation experiment results of expert system in the heating furnace show that the inference speed is fast and the result is satisfactory.