进化粗糙集的逼近最优性
来源期刊:中南大学学报(自然科学版)2000年第1期
论文作者:刘健勤
文章页码:81 - 83
关键词:粗糙集;进化计算;优化理论
Key words:rough set; evolutionary computation; optimization theory
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,在非确定信息建模与非线性信号处理方面具有重要的意义.
Abstract: 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 setmechanism). The paper proposes and proves the theorems of evolutionaryrough logicwithmartingle process and global optimality condition with corresponding quantitatve description. Results show that the evolutionary rough set constructed gives the necessarytheory basis and formal method for the solution process of the generalized problemwhich combines the nonmonotonous logic of character and nonlinear progamming of numerical value. It is important in modelling of uncertain information and processing nonlinear signals.