进化粗糙集的逼近最优性  

Approximation optimality for evolutionary rough set

作　　者：刘健勤[1] 

机构地区：[1]中南工业大学信息科学与工程学院,湖南长沙410083

出　　处：《中南工业大学学报》2000年第1期81-83,共3页Journal of Central South University of Technology(Natural Science)

基　　金：国家自然科学基金资助项目! ( 699740 43 );国家教育部留学回国人员科研基金资助项目 !( 960 7EZ0 4)

摘　　要：构造了进化粗糙集的表征测度和分解形式 ,讨论了其随机过程和信息几何的属性 ,提出并证明了逼近求解方面其鞅过程和全局最优性条件的数学结论 ,同时给出了相关的定量描述 .结果表明 ,所建立的进化粗糙集结构为符号的非单调逻辑与数值的非线性规划相统一的广义问题求解过程设计提供了必需的理论基础和形式化手段 。This paper constructs the representation measurement and decomposition form for evolutionary rough set and discusses the characteristics with respect to its random processes and information geometry (concerned in evolutionary rough set mechanism). The paper proposes and proves the theorems of evolutionary rough logic with martingle process and global optimality condition with corresponding quantitatve description. Results show that the evolutionary rough set constructed gives the necessary theory basis and formal method for the solution process of the generalized problem which combines the nonmonotonous logic of character and nonlinear progamming of numerical value. It is important in modelling of uncertain information and processing nonlinear signals. [

关 键 词：粗糙集 进化计算 优化理论 逼近最优性 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]