大跨空间结构智能选型方案生成系统  

Scheme generation system for intelligent form selection of large-span spatial structure

作　　者：王力[1] 吕大刚[1] 刘晓燕[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》2004年第11期1435-1438,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);国家自然科学基金资助项目(50378030);黑龙江省自然科学基金资助项目(E00-01).

摘　　要：将基于实例的推理和数据挖掘技术应用于大跨空间结构的智能选型的方案生成阶段.介绍了基于实例推理系统的基本结构及其四个关键技术,建造了大跨空间结构智能选型方案生成系统.主要实现了以下功能:实例库的建立和维护;人机交互界面的实现;初步完成基于加权检索算法的大跨空间结构选型实例推理;实现了实例关键参数值域分布情况的统计功能;采用数据挖掘算法对实例库进行了关联规则的挖掘,初步挖掘出了实例库中的关联规则,利用决策树和粗糙集的方法产生分类规则,比较了三种数据挖掘方法,给出它们的应用条件,为下一步实现基于知识推理做准备.通过实例演示了该系统的方案生成过程.Case-based reasoning (CBR) together with data mining (DM) technique is applied in intelligent generation system of large-span spatial structural schemes. The basic system structure of case-based reasoning and its four key techniques are firstly presented. Combining CBR with DM, an intelligent scheme generation system of large-span spatial structures is developed. The following main functions, such as, a case-base of large-span spatial structural schemes and its maintenance module, a man-machine interactive interface, a weighted retrieving algorithms for CBR, the association rules imbed in case-base using data mining algorithms and the classification rules implicit in case-base using decision trees and rough sets, are realized. Three data mining methods are compared, and then their application potentials are given in order to prepare for the further knowledge-based reasoning. Finally, the scheme generation procession using the developed system is shown by a practical engineering example.

关 键 词：大跨空间结构 方案生成 实例推理 数据挖掘 粗糙集 

分 类 号：TU318[建筑科学—结构工程]