基于神经模式识别的建筑结构选型方案研究  

Scheme generation system for structural form selection based on neural pattern recognition

作　　者：吕大刚[1] 邢方亮[1] 张世海[1] 王光远[1] 

机构地区：[1]哈尔滨工业大学土木工程学院,黑龙江哈尔滨150090

出　　处：《哈尔滨工业大学学报》2003年第12期1418-1421,共4页Journal of Harbin Institute of Technology

基　　金：国家自然科学基金资助重大项目(59895410);黑龙江省自然科学基金资助项目(E00-01)

摘　　要：提出了基于实例推理的建筑结构智能选型模型,揭示了实例检索的模式识别本质。提出了基于神经识别的实例推理方法,并将其应用于建筑结构智能选型的方案生成阶段,通过对高层建筑结构的实际选型设计证明了本方法的有效性。An intelligent design model of structural form selection based on case reasoning is put forward to simulate the way of thinking in images of experts. The essence of case retrieval lies in clustering and classification of patterns. Neural computing, pattern recognition and case-based reasoning (CBR) are integrated into a hybrid intelligent design system, to generate a CBR method based on neural pattern recognition. The effectiveness of this method is shown by its application in the design practice of structural form selection of high-rise buildings.

关 键 词：建筑结构 智能选型 神经模式识别 距离函数 

分 类 号：TU318[建筑科学—结构工程]