基于粗糙集正域决策树的结构智能型式优化知识发现  

STRUCTURAL INTELLIGENT PATTERN OPTIMIZATION KNOWLEDGE DISCOVERY BASED ON ROUGH SET POSITIVE DOMAIN DECISION TREE

作　　者：张远征 ZHANG Yuan-zheng(Department of Computer Science,University of Southern California,Los Angeles 90007,USA)

机构地区：[1]南加州大学计算机科学系,美国洛杉矶90007

出　　处：《南阳理工学院学报》2025年第2期53-58,共6页Journal of Nanyang Institute of Technology

摘　　要：结构智能型式优化是工程结构优化领域未被很好解决的难题,知识获取是其“瓶颈”和关键。首先,分析了基于决策树、粗糙集与智能算法的知识发现方法的特征与不足,指出决策树方法可较好克服后者直观性与可解释性不强问题,粗糙集可弥补决策树优化难度大问题;接着,利用粗糙集正域能充分考虑条件与决策属性间依赖性及计算简洁的特征,构建了一种基于粗糙集正域决策树的知识发现新方法;最后,给出了其在高层建筑结构智能型式优化知识获取中的应用。实践表明,新方法兼有二者的共同优点,拓展了传统决策树方法,提高了决策树的构建效率,为解决结构智能型式优化及智能知识发现方法可解释性问题提供了新路径。Structure intelligent pattern optimization is an unsolved problem in the field of engineering structure optimization,and knowledge acquisition is its″bottleneck″and key.Firstly,the characteristics and shortcomings of knowledge discovery methods based on decision tree,rough set and intelligent algorithm are analyzed,and it is pointed out that decision tree method can overcome the intuitiveness and interpretability of the latter,and rough set can make up for the difficulty of decision tree optimization.Then,a new method of knowledge discovery based on rough set positive domain decision tree is constructed by making use of the characteristic that the dependence between conditions and decision attributes can be fully considered and the calculation is concise.Finally,its application in the knowledge acquisition of intelligent type optimization of high-rise building structure is given.Practice shows that the new method has the advantages of both,expands the traditional decision tree method,improves the efficiency of decision tree construction,and provides a new way to solve the problems of structural intelligent pattern optimization and the interpretability of intelligent knowledge discovery method.

关 键 词：粗糙集 正域 决策树 结构型式 智能优化 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程] TU973.1[自动化与计算机技术—控制科学与工程]