三角范数的构造 T-norm 和三角余范数 T-conorm

Fuzzy set theory

crossover point

令模糊集\(A\)具有隶属函数\(\mu_{A}(x)\),该隶属函数定义域\(U\)\(A\)的交叉点定义为

\[crossover(A)=\{x\in U | \mu_{A}(x)=0.5\} \]

\(\alpha\)-cut

令模糊集\(A\)具有隶属度函数\(\mu_{A}(x)\)\(\mu_{A}(x)\)的定义域为\(U\)。则\(A\)\(\alpha\)-cut定义为

\[A_{\alpha}=\{x\in U| \,u_{A}(x)\geq \alpha\} \]

其中\(\alpha\)-cut定义为\(\alpha \in [0,1]\)

投影projection

令模糊集\(A\)具有隶属函数\(\mu_{A}(x)\)\(\mu_{A}(x)\)的定义域\(U\)\(R^{n}\)的子集,亦即\(x=(x_{1},x_{2},\ldots,x_{n})\in R^{n}\)。这里\(x\)为n-dim的vector,向量。

\(x_{1}=0\)代表\(x_{2},x_{3}\ldots x_{n}\)的子空间

\[S_{1}=''x_{1}=0''=\{ (0,x_{2},\ldots,x_{n})|x_{i}\in R \} \\= \{ (x_{1},x_{2},\ldots,x_{n})\in R^{n} |x_{1}=0\} \]

\(A\)\(S_{1}\)的投影为

\[\mu_{A_{S_{1}}}(x_{2},\ldots,x_{n})=sup_{x_{1}\in R} \mu_{A}(x_{1},x_{2},\ldots,x_{n}) \]

这里\(x_{1}\)为free variable。\(sup(A)\) denotes the support of fuzzy set \(A\):

\[sup(A) = \{ x\in U| \mu_{A}(x)>0 \} \]

Cartesian Product

两个集合\(U\)\(V\)形成:

\[U\times V = \{ (u,v)|u\in U,v \in V \} \]

其中\((u,v)\)是有序对
\(U\neq V\),则\(U\times V \neq V\times U\)
\(U=V\),则\(V\times V = V^{2}\)

\[R^{2} = \{ (u,v)|u,v\in R \} \]

例如
\(U=\{ 1,2,3\}, V=\{ 4,5\}\)
\(U\times V =\{ (1,4),(1,5),(2,4),(2,5),(3,4),(3,5) \}\)

A variable can take words as its values. We call this variable as a linguistic variable.

  1. Words大致引用自然语言。
  2. Words必须具有可被运算性或可被判断的数学模式,主要是可以以模糊集形式呈现出来。

Temperature is high. Temperature is linguistic variable. high is a fuzzy set.

语言变量的正式定义

\((X,T,U,M)\)
\(X\)linguistic variable 的名称集合
\(T\)linguistic variable 可以选取的linguistic value 之集合。
\(U\) actual physical doman
\(M\)描绘\(T\)linguistic variable对应的模糊函数的语义规则

模糊命题一般有两种
原型模糊命题(atomic fuzzy propositions)
x is A.
合成模糊命题(compound fuzzy propositions)
\(x_{1}\) is \(A_{1}\) and \(x_{2}\) is \(A_{2}\) or \(x_{3}\) is \(A_{3}\)

Convolution

Logic (s-t composition) of convolution

T-norm

Pooling

t-conorms

三角范数构造

参考 --‘’三角范数构造‘’ 薛占军,刘三阳,西北大学学报2009.4.39卷2期

三角范数是取值于[0,1]的二元函数,因此三角范数在概率度量空间、决策论、统计学、博弈、函数方程等领域有这重要的应用价值。
定义1 -- 设\(T:[0,1]^2 \rightarrow[0,1]\)是[0,1]上的二元运算,若对任意

posted @ 2021-09-14 20:31  温酒待君归  阅读(1020)  评论(0编辑  收藏  举报
levels of contents