高级智能的数学基础
1.习题2.2.1
设 D ⊂ R n D\sub \bf R^n D⊂Rn是有限集,则方程组(2.2.5)等价与
y = s g n ( ( w , x ) ) = { + 1 , x ∈ A − 1 , x ∈ B (2.2.5) y=sgn((w,x))=\begin{cases} +1 ,\qquad x \in A\\ -1 ,\qquad x\in B\end{cases} \tag{2.2.5} y=sgn((w,x))={+1,x∈A−1,x∈B(2.2.5)
( w , x ) = ∑ j = 1 n w j x j > 0 , ∀ x ∈ D . (2.2.6) (w,x)= \sum_{j=1}^n w_jx_j>0,\qquad \forall x\in D. \tag{2.2.6} (w,x)=j=1∑nwjxj>0,∀x∈D.(2.2.6)
其中: A ∩ B = ∅ D = A ∪ ( − B ) ( − B ) = { − x ∣ x ∈ B } A \cap B=\varnothing \qquad D=A \cup(-B) \qquad (-B)=\{-x|x \in B \} A∩B=∅D=A∪(−B)(−B)={−x∣x∈B} ,
证明:(1)必要性:对于 ∀ x ∈ D \forall x \in D ∀x∈D,当 x ∈ A x\in A x∈A时, s g n ( ( w , x ) ) = + 1 ( w , x ) = ∑ j = 1 n w j x j > 0 sgn((w,x))=+1 \qquad (w,x)=\sum\limits_{j=1}^n w_jx_j>0 sgn((w,x))=+1(w,x)=j=1∑nwjxj>0
当 x ∈ B x\in B x∈B时, s g n ( ( w , x ) ) = − 1 sgn((w,x))=-1 sgn((w,x))=−1 , ( w , x ) < 0 (w,x)<0 (w,x)<0,则有 ( w , − x ) > 0 (w,-x)>0 (w,−x)>0, 则 x ∈ ( − B ) x \in(-B) x∈(−B)时, ( w , x ) > 0 (w,x)>0 (w,x)>0
所以,方程组
y
=
sgn
(
(
w
,
x
)
)
=
{
+
1
,
x
∈
A
−
1
,
x
∈
B
y=\operatorname{sgn}((w, x))=\left\{\begin{array}{ll} +1, & x \in A \\ -1, & x \in B \end{array}\right.
y=sgn((w,x))={+1,−1,x∈Ax∈B
可以得到
(
w
,
x
)
=
∑
j
=
1
n
w
j
x
j
>
0
,
∀
x
∈
D
.
(w,x)= \sum_{j=1}^n w_jx_j>0,\qquad \forall x\in D.
(w,x)=∑j=1nwjxj>0,∀x∈D.
(2) 充分性:对于 ( w , x ) = ∑ j = 1 n w j x j > 0 , ∀ x ∈ D , 其中 D = A ∪ ( − B ) ,且 A ∩ B = ∅ (w,x)= \sum_{j=1}^nw_jx_j>0,\qquad \forall x\in D,其中D=A \cup(-B),且A \cap B=\varnothing (w,x)=∑j=1nwjxj>0,∀x∈D,其中D=A∪(−B),且A∩B=∅
当 x ∈ A 时,( w , x ) = ∑ j = 1 n w j x j > 0 , s g n ( ( w , x ) ) = 1 当x \in A时,(w,x)=\sum_{j=1}^n w_jx_j>0,\qquad sgn((w,x))=1 当x∈A时,(w,x)=∑j=1nwjxj>0,sgn((w,x))=1
当 x ∈ ( − B ) 时,( w , x ) = ∑ j = 1 n w j x j > 0 , ∀ x ∈ D 可得到 y = s g n ( ( w , x ) ) = { + 1 , x ∈ A − 1 , x ∈ B 当x\in(-B)时,(w,x)=\sum_ {j=1}^n w_jx_j>0 ,\quad \forall x\in D 可得到y=sgn((w,x))=\begin{cases} +1 ,\qquad x \in A\\ -1 ,\qquad x\in B\end{cases} 当x∈(−B)时,(w,x)=∑j=1nwjxj>0,∀x∈D可得到y=sgn((w,x))={+1,x∈A−1,x∈B
所以,由 ( 1 ) ( 2 ) 得 y = s g n ( ( w , x ) ) = = { + 1 , x ∈ A − 1 , x ∈ B , 等价与 ( w , x ) = ∑ j = 1 n w j x j > 0 , ∀ x ∈ D 所以,由(1)(2)得y=sgn((w,x))==\begin{cases} +1 ,\qquad x \in A\\ -1 ,\qquad x\in B\end{cases},\quad 等价与(w,x)=\sum\limits_{j=1}^nw_jx_j>0,\quad \forall x \in D 所以,由(1)(2)得y=sgn((w,x))=={+1,x∈A−1,x∈B,等价与(w,x)=j=1∑nwjxj>0,∀x∈D。
2. 习题3.5
设二层前向网中激励函数为双曲正切函数,讨论其学习算法。
解:设二层前向网中激励函数为
φ
(
u
)
=
φ
′
(
u
)
=
2
1
+
e
−
2
u
−
1
\varphi(u)= \varphi {}' (u) = \frac{2}{1+e^{-2u}} -1
φ(u)=φ′(u)=1+e−2u2−1
并取
E
=
1
2
∑
α
=
1
N
∑
h
=
1
c
(
y
h
α
−
t
h
α
)
2
E = \frac{1}{2} \sum \limits ^{N} _{\alpha =1} \sum \limits ^{c} _{h=1} (y^{\alpha}_h -t^{\alpha}_h) ^2
E=21α=1∑Nh=1∑c(yhα−thα)2 ,其中记
E
α
=
1
2
∑
h
=
1
c
(
y
h
α
−
t
h
α
)
2
E^{\alpha} = \frac{1}{2} \sum \limits ^{c} _{h=1} (y^{\alpha}_h -t^{\alpha}_h) ^2
Eα=21h=1∑c(yhα−thα)2
其中 φ ′ ( u ) = 1 − φ 2 ( u ) \varphi^{'}(u) = 1-\varphi^{2}(u) φ′(u)=1−φ2(u)
于是
δ
k
α
(
2
)
=
[
1
−
φ
(
a
~
k
α
)
2
]
(
y
k
α
−
t
k
α
)
δ
j
α
(
1
)
=
[
1
−
φ
(
a
~
j
α
)
2
]
∑
k
=
1
c
w
k
j
(
2
)
δ
k
α
(
2
)
\delta ^{(2)} _{k\alpha } =[1-\varphi(\tilde{a}^{\alpha}_k )^2 ](y^{\alpha }_k-t^{\alpha }_k) \\ \delta ^{(1)} _{j\alpha } =[1-\varphi(\tilde{a}^{\alpha}_j )^2 ]\sum ^{c} _{k=1} w^{(2)}_{kj} \delta ^{(2)}_{k\alpha }
δkα(2)=[1−φ(a~kα)2](ykα−tkα)δjα(1)=[1−φ(a~jα)2]k=1∑cwkj(2)δkα(2)
其中
a
j
α
=
∑
i
=
0
d
w
j
i
1
x
i
α
a^{\alpha } _j = \sum \limits ^d _{i=0}w^{1}_{ji}x^{\alpha }_{i}
ajα=i=0∑dwji1xiα ,
z
j
α
=
φ
(
a
j
α
)
z^{\alpha }_{j} = \varphi (a^{\alpha }_{j})
zjα=φ(ajα) ,
a
~
k
α
=
∑
j
=
0
M
w
k
j
2
z
j
α
\tilde{a} ^{\alpha }_k = \sum \limits ^M _{j=0}w^{2}_{kj}z^{\alpha }_{j}
a~kα=j=0∑Mwkj2zjα ,
y
k
α
=
φ
(
a
~
k
α
)
y^{\alpha } _k = \varphi(\tilde{a} ^{\alpha }_k)
ykα=φ(a~kα)
迭代过程,计算第
t
+
1
t+1
t+1 次权重
w
k
j
(
2
)
(
t
+
1
)
=
w
k
j
(
2
)
(
t
)
−
η
∑
α
=
1
N
δ
k
α
(
2
)
(
t
)
⋅
z
j
α
w
j
i
(
1
)
(
t
+
1
)
=
w
j
i
(
1
)
(
t
)
−
η
∑
α
=
1
N
δ
j
α
(
1
)
(
t
)
⋅
x
i
α
w^{(2)}_{kj}(t+1) = w^{(2)}_{kj}(t)-\eta \sum ^{N} _{\alpha =1} \delta ^{(2)}_{k\alpha }(t)\cdot z^{\alpha }_j \\ w^{(1)}_{ji}(t+1) = w^{(1)}_{ji}(t)-\eta \sum ^{N} _{\alpha =1} \delta ^{(1)}_{j\alpha }(t)\cdot x^{\alpha }_i
wkj(2)(t+1)=wkj(2)(t)−ηα=1∑Nδkα(2)(t)⋅zjαwji(1)(t+1)=wji(1)(t)−ηα=1∑Nδjα(1)(t)⋅xiα
由上述过程可得序列
{
δ
j
α
(
1
)
(
t
)
,
δ
k
α
(
2
)
(
t
)
,
w
j
i
(
1
)
(
t
)
,
w
k
j
(
2
)
(
t
)
}
\left \{ \delta ^{(1)}_{j\alpha } (t), \delta ^{(2)}_{k\alpha } (t), w^{(1)}_{ji } (t),w^{(2)}_{kj } (t) \right \}
{δjα(1)(t),δkα(2)(t),wji(1)(t),wkj(2)(t)} , 直至满足停机准则。
解:tanh是双曲函数中的一个,tanh() 为双曲正切。在数学中,双曲正切 tanh 是由双曲正弦和双曲余弦这两者基本双曲函数推导而来。
正切函数时非常常见的激活函数,与Sigmoid函数相比,它的输出均值是0,使得其收敛速度要比Sigmoid快,减少迭代次数。相对于
Sigmoid的好处是它的输出的均值为0,克服了第二点缺点。但是当饱和的时候还是会杀死梯度。
3. Bp算法
给定样本为 x = [1, -1, 1] , y =[1, 1],选 η = 0.1 \eta =0.1 η=0.1,计算下图所示网络的权值调节过程(用BP算法)。 隐单元及输出单元用 φ ( x ) = 2 1 + e − x − 1 \varphi(x)=\frac 2 {1+e^{-x}}-1 φ(x)=1+e−x2−1 作为激活函数,初始权值选为
w 1 = [ 1 1 1 2 0 2 3 3 − 3 ] w 2 = [ 1 − 1 1 0 1 1 ] w_1= \begin{bmatrix} 1&1&1\\ 2&0&2\\ 3&3&-3 \end{bmatrix} \qquad \qquad \qquad w_2= \begin{bmatrix} 1&-1&1\\ 0&1&1 \end{bmatrix} w1= 12310312−3 w2=[10−1111]
解:(1)设神经元输出用O表示,其编号如图所示,则隐含层和输出层输出为 O j = f ( n e t j ) = 2 1 + e − n e t i − 1 O_j=f(net_j)=\frac 2 {1+e^{-net_i}}-1 Oj=f(netj)=1+e−neti2−1 , 激活函数求导为 f ′ ( x ) = 1 2 [ 1 − f 2 ( x ) ] = 1 2 ( 1 − O j 2 ) f^{\prime}(x)=\frac 1 2\left[1-f^2(x)\right]=\frac 1 2(1-O_j^2) f′(x)=21[1−f2(x)]=21(1−Oj2) 。
(2)前向传播
隐含层: n e t 4 = x 1 w 11 + x 2 w 12 + x 3 w 13 = 1 ∗ 1 + 1 ∗ ( − 1 ) + 1 ∗ 1 = 1 net_4=x_1w_{11}+x_2w_{12}+x_3w_{13}=1*1+1*(-1)+1*1=1 net4=x1w11+x2w12+x3w13=1∗1+1∗(−1)+1∗1=1
n e t 5 = x 1 w 21 + x 2 w 22 + x 3 w 23 = 2 ∗ 1 + 0 ∗ ( − 1 ) + 2 ∗ 1 = 4 net_5=x_1w_{21}+x_2w_{22}+x_3w_{23}=2*1+0*(-1)+2*1=4 net5=x1w21+x2w22+x3w23=2∗1+0∗(−1)+2∗1=4
n e t 6 = x 1 w 31 + x 2 w 32 + x 3 w 33 = 3 ∗ 1 + 3 ∗ ( − 1 ) + ( − 3 ) ∗ 1 = − 3 net_6=x_1w_{31}+x_2w_{32}+x_3w_{33}=3*1+3*(-1)+(-3)*1=-3 net6=x1w31+x2w32+x3w33=3∗1+3∗(−1)+(−3)∗1=−3
O 4 = f ( n e t 4 ) = 0.4621 O_4=f(net_4)=0.4621 O4=f(net4)=0.4621
O 5 = f ( n e t 5 ) = 0.964 O_5=f(net_5)=0.964 O5=f(net5)=0.964
O 6 = f ( n e t 6 ) = − 0.9051 O_6=f(net_6)=-0.9051 O6=f(net6)=−0.9051
输出层: n e t y 1 = O 4 w 11 + O 5 w 12 + O 6 w 13 = 0.4621 ∗ 1 + 0.964 ∗ ( − 1 ) + ( − 0.9051 ) ∗ 1 = − 1.407 nety_1=O_4w_{11}+O_5w_{12}+O_6w_{13}=0.4621*1+0.964*(-1)+(-0.9051)*1=-1.407 nety1=O4w11+O5w12+O6w13=0.4621∗1+0.964∗(−1)+(−0.9051)∗1=−1.407
n e t y 2 = O 4 w 21 + O 5 w 22 + O 6 w 23 = 0.4621 ∗ 0 + 0.964 ∗ 1 + ( − 0.9051 ) ∗ 1 = 0.0589 nety_2=O_4w_{21}+O_5w_{22}+O_6w_{23}=0.4621*0+0.964*1+(-0.9051)*1=0.0589 nety2=O4w21+O5w22+O6w23=0.4621∗0+0.964∗1+(−0.9051)∗1=0.0589
O y 1 = f ( n e t y 1 ) = − 0.6065 O_{y_1}=f(net_{y_1})=-0.6065 Oy1=f(nety1)=−0.6065
O y 2 = f ( n e t y 2 ) = 0.0294 O_{y_2}=f(net_{y_2})=0.0294 Oy2=f(nety2)=0.0294
(3)构造损失函数
E = 1 2 ∑ k = 1 N ( y k − y k ^ ) 2 E=\frac 1 2 \sum_ {k=1}^N(y_k-\hat{y_k})^2 E=21∑k=1N(yk−yk^)2, 其中 E k = ( y k − y k ^ ) 2 E_k=(y_k-\hat{y_k})^2 Ek=(yk−yk^)2
(4)用反向传播求解偏导数
E t o t a l = E y 1 + E y 2 = 1 2 ( 1 − ( − 0.6065 ) 2 ) + 1 2 ( 1 − ( 0.0294 ) 2 ) = 1.76145 E_{total} = E_{y_1} + E_{y_2} = \frac 1 2 (1-(-0.6065)^2)+ \frac 1 2 (1-(0.0294)^2) = 1.76145 Etotal=Ey1+Ey2=21(1−(−0.6065)2)+21(1−(0.0294)2)=1.76145
根据公式
∂
E
∂
w
=
∂
E
∂
o
y
⋅
∂
o
y
∂
n
e
t
y
⋅
∂
n
e
t
y
∂
w
\frac{\partial E}{\partial w} =\frac{\partial E}{\partial o_y} \cdot \frac{\partial o_y}{\partial net_y} \cdot \frac{\partial net_y}{\partial w}
∂w∂E=∂oy∂E⋅∂nety∂oy⋅∂w∂nety
权值更新公式:
w
+
=
w
−
η
∂
E
∂
w
w^+ = w - \eta \frac{\partial E}{\partial w}
w+=w−η∂w∂E
带入数据得到:
w 1 = [ 1.03 0.97 1.030 1.99 0.01 1.99 3.041 2.956 − 2.958 ] w 2 = [ 1.035 0.93 0.94 0.212 1.037 0.964 ] w_1=\begin{bmatrix} 1.03& 0.97 & 1.030\\ 1.99& 0.01 & 1.99\\ 3.041& 2.956 & -2.958 \end{bmatrix} \quad \quad\quad\quad w_2 = \begin{bmatrix} 1.035 & 0.93 & 0.94\\ 0.212 & 1.037 & 0.964 \end{bmatrix} w1= 1.031.993.0410.970.012.9561.0301.99−2.958 w2=[1.0350.2120.931.0370.940.964]
4.命题3.5.2
设圈 C 1 , C 2 , . . . , C k 是图 G 中的全体图,那么 D ( C 1 ) , D ( C 2 ) , . . . , D ( C k ) 两两不交,且并集为 A , 称之为空间 A 的一个划分 设圈C_1,C_2,...,C_k是图G中的全体图,那么D(C_1),D(C_2),...,D(C_k)两两不交,且并集为A,\\称之为空间A的一个划分 设圈C1,C2,...,Ck是图G中的全体图,那么D(C1),D(C2),...,D(Ck)两两不交,且并集为A,称之为空间A的一个划分
**证明:图G默认为是由HNNS生成的图,则图G中每一个点有且只有一条出弧,图G中不同的圈无公共点,且图G中每个点通过若干个点必可达到一个圈。题意中D©为全体可到圈C的点,因为图G中不同的圈无公共点,且图G中每个点通过若干个点必可到一个圈,故命题得证。
5.数值实验
在平面 R 2 上取矩形 I = [ − 5 , 5 ] × ⌊ − 5 , 5 ⌋ , 并在 I 中随机产生 100 个样本点,作为训练样本点,得到训练样本 集 Ω = { ( x i , y i ) ∣ i = 1 , 2 , . . . , 100 } ⊂ I × { − 1 , + 1 } , 其中,如果 x i 落在圆 t 1 2 + t 2 2 = 16 ( ( t 1 , t 2 ) ) ∈ R 2 内, 则取 y i = − 1 , 否则取 y i = + 1 \hspace{0cm} 在平面R^2 上取矩形I=[-5,5] \times \lfloor -5,5 \rfloor ,并在I中随机产生100个样本点,作为训练样本点,得到训练样本\\集 \Omega =\{(x_i,y_i)|i=1,2,...,100 \} \subset I \times \{-1,+1 \} ,其中,如果x_i落在圆t^2_1+t^2_2=16 ((t_1,t_2)) \in \bf R^2 内,\\则取y_i=-1,否则取y_i=+1 在平面R2上取矩形I=[−5,5]×⌊−5,5⌋,并在I中随机产生100个样本点,作为训练样本点,得到训练样本集Ω={(xi,yi)∣i=1,2,...,100}⊂I×{−1,+1},其中,如果xi落在圆t12+t22=16((t1,t2))∈R2内,则取yi=−1,否则取yi=+1
采用Gauss支撑向量机进行学习。Gauss核函数的参数选取及其分类效果,见表4.4.1.学习后的得到的分类曲线见图4.4.1
6.习题 8
8.1 设
X
=
{
1
,
2
,
3
,
4
,
5
,
6
,
7
}
,
X=\{1,2,3,4,5,6,7\},
X={1,2,3,4,5,6,7},
A
∈
F
(
X
)
,
A \in F(X),
A∈F(X), 其隶属度
A
~
(
x
)
\tilde{A}(x)
A~(x) 如下:
A
~
(
1
)
=
0.1
A
~
(
2
)
=
0.3
A
~
(
3
)
=
0.8
A
~
(
4
)
=
1
A
~
(
5
)
=
0.8
A
~
(
6
)
=
0.3
A
~
(
7
)
=
0
\tilde{A}(1)=0.1 \quad \tilde{A}(2)=0.3 \quad \tilde{A}(3)=0.8 \quad \tilde{A}(4)=1 \\ \tilde{A}(5)=0.8 \quad \tilde{A}(6)=0.3 \quad \tilde{A}(7)=0
A~(1)=0.1A~(2)=0.3A~(3)=0.8A~(4)=1A~(5)=0.8A~(6)=0.3A~(7)=0
- 分别用查德法,向量法、序偶法表示 A ~ \tilde{A} A~ ;
- 求 A c ~ \tilde{A^c} Ac~ ;
- 指出 A ~ \tilde{A} A~ 的意义.
【解】: (1)查德法: A ~ = 0.1 1 + 0.3 2 + 0.8 3 + 1 4 + 0.8 5 + 0.3 6 + 0 7 \tilde{A} = \frac {0.1} {1} +\frac {0.3} {2} +\frac {0.8} {3} +\frac {1} {4} +\frac {0.8} {5} +\frac {0.3} {6} +\frac {0} {7} A~=10.1+20.3+30.8+41+50.8+60.3+70
向量法: A ~ = ( 0.1 , 0.3 , 0.8 , 1 , 0.8 , 0.3 , 0 ) \tilde{A}=(0.1,0.3,0.8,1,0.8,0.3,0) A~=(0.1,0.3,0.8,1,0.8,0.3,0)
序偶法: A ~ = { ( 1 , 0.1 ) , ( 2 , 0.3 ) , ( 3 , 0.8 ) , ( 4 , 1 ) , ( 5 , 0.8 ) , ( 6 , 0.3 ) , ( 7 , 0 ) } \tilde{A}=\{(1,0.1),(2,0.3),(3,0.8),(4,1),(5,0.8),(6,0.3),(7,0)\} A~={(1,0.1),(2,0.3),(3,0.8),(4,1),(5,0.8),(6,0.3),(7,0)}
(2)
A
c
~
=
1
−
0.1
1
+
1
−
0.3
2
+
1
−
0.8
3
+
1
−
1
4
+
1
−
0.8
5
+
1
−
0.3
6
+
1
−
0
7
=
0.9
1
+
0.7
2
+
0.2
3
+
0
4
+
0.2
5
+
0.7
6
+
1
7
\begin{align} \tilde{A^c}&=\frac {1-0.1} 1 +\frac {1-0.3} 2 +\frac {1-0.8} 3 +\frac {1-1} 4 +\frac {1-0.8} 5 +\frac {1-0.3} 6 +\frac {1-0} 7 \\ &= \frac {0.9} 1 + \frac {0.7} 2 +\frac {0.2} 3 +\frac {0 } 4 +\frac {0.2} 5 +\frac {0.7} 6 +\frac {1} 7 \end{align}
Ac~=11−0.1+21−0.3+31−0.8+41−1+51−0.8+61−0.3+71−0=10.9+20.7+30.2+40+50.2+60.7+71
(3)
A
~
\tilde{A}
A~ 表示在集合
X
X
X 中的元素,在
A
~
(
x
)
\tilde{A}(x)
A~(x) 的映射下有一个实数值
A
~
(
x
)
∈
[
0
,
1
]
\tilde{A}(x)\in [0,1]
A~(x)∈[0,1] 与之对应,则
A
~
\tilde{A}
A~ 表示元素
x
x
x 属于集合
A
A
A 的程度。
8.2已知模糊集“老年” 和 “年轻” 的隶属函数分别为:
O
~
(
x
)
=
{
0
,
当
0
⩽
x
⩽
50
时,
[
1
+
(
x
−
50
5
)
−
2
]
−
1
,
当
50
<
x
⩽
200
时,
Y
~
(
x
)
=
{
1
,
当
0
⩽
x
⩽
25
时,
[
1
+
(
x
−
25
5
)
2
]
−
1
,
当
25
<
x
⩽
200
时.
\begin{array}{l} \widetilde{O}(x)=\left\{\begin{array}{ll} 0, & \text { 当 } 0 \leqslant x \leqslant 50 \text { 时, } \\ {\left[1+\left(\frac{x-50}{5}\right)^{-2}\right]^{-1},} & \text { 当 } 50<x \leqslant 200 \text { 时, } \end{array}\right. \\ \tilde{Y}(x)=\left\{\begin{array}{ll} 1, & \text { 当 } 0 \leqslant x \leqslant 25 \text { 时, } \\ {\left[1+\left(\frac{x-25}{5}\right)^{2}\right]^{-1},} & \text { 当 } 25<x \leqslant 200 \text { 时. } \end{array}\right. \\ \end{array}
O
(x)={0,[1+(5x−50)−2]−1, 当 0⩽x⩽50 时, 当 50<x⩽200 时, Y~(x)={1,[1+(5x−25)2]−1, 当 0⩽x⩽25 时, 当 25<x⩽200 时.
试写出模糊集“不老” 和 “既不老又不年轻” 的隶属函数.
【解】:“不老”的隶属函数:
[
O
~
(
x
)
]
c
=
{
1
,
当
0
⩽
x
⩽
50
时,
[
1
+
(
x
−
50
5
)
2
]
−
1
当
50
<
x
⩽
200
时,
[\widetilde{O}(x)]^c=\left\{ \begin{array}{ll}1, & \text { 当 } 0 \leqslant x \leqslant 50 \text { 时, } \\ {\left[1+\left(\frac{x-50}{5}\right)^{2}\right]^{-1}} & \text { 当 } 50<x \leqslant 200 \text { 时, } \end{array}\right.
[O
(x)]c={1,[1+(5x−50)2]−1 当 0⩽x⩽50 时, 当 50<x⩽200 时,
“不年轻”的隶属函数:
[
Y
~
(
x
)
]
c
=
{
0
,
当
0
⩽
x
⩽
25
时,
[
1
+
(
x
−
25
5
)
−
2
]
−
1
当
25
<
x
⩽
200
时,
[\widetilde{Y}(x)]^c=\left\{ \begin{array}{ll}0, & \text { 当 } 0 \leqslant x \leqslant 25 \text { 时, } \\ {\left[1+\left(\frac{x-25}{5}\right)^{-2}\right]^{-1}} & \text { 当 } 25<x \leqslant 200 \text { 时, } \end{array}\right.
[Y
(x)]c={0,[1+(5x−25)−2]−1 当 0⩽x⩽25 时, 当 25<x⩽200 时,
“既不老又不年轻”的隶属函数:
[
O
~
(
x
)
]
c
∩
[
Y
~
(
x
)
]
c
=
{
0
,
当
0
⩽
x
⩽
25
时,
[
1
+
(
x
−
25
5
)
−
2
]
−
1
当
25
<
x
⩽
51
时,
[
1
+
(
x
−
50
5
)
2
]
−
1
当
51
<
x
⩽
200
时,
[\widetilde{O}(x)]^c \cap [\widetilde{Y}(x)]^c=\left\{ \begin{array}{ll}0, & \text { 当 } 0 \leqslant x \leqslant 25 \text { 时, } \\ {\left[1+\left(\frac{x-25}{5}\right)^{-2}\right]^{-1}} & \text { 当 } 25<x \leqslant 51\text { 时, }\\ {\left[1+\left(\frac{x-50}{5}\right)^{2}\right]^{-1}} & \text { 当 } 51<x \leqslant 200 \text { 时, } \end{array}\right.
[O
(x)]c∩[Y
(x)]c=⎩
⎨
⎧0,[1+(5x−25)−2]−1[1+(5x−50)2]−1 当 0⩽x⩽25 时, 当 25<x⩽51 时, 当 51<x⩽200 时,
8.3 设 A ~ , B ~ , C ~ ∈ F ( x ) \tilde{A}, \tilde{B},\tilde{C} \in F(x) A~,B~,C~∈F(x),如下表:
x 1 x_1 x1 | x 2 x_2 x2 | x 3 x_3 x3 | x 4 x_4 x4 | x 5 x_5 x5 | x 6 x_6 x6 | |
---|---|---|---|---|---|---|
A ~ \tilde{A} A~ | 0.1 | 0.6 | 0.8 | 0.9 | 0.7 | 0.1 |
B ~ \tilde{B} B~ | 0.9 | 0.7 | 0.5 | 0.2 | 0.1 | 0 |
C ~ \tilde{C} C~ | 0.6 | 0.8 | 1 | 0.7 | 0.5 | 0.3 |
求: A ~ ∪ B ~ , A ~ ∩ B ~ , ( A ~ ∪ B ~ ) c , ( A ~ ∩ B ~ ) c , ( A ~ ∪ B ~ ) c ∩ C ~ , ( A ~ ∪ B ~ ) c ∪ C ~ , ( A ~ ∩ B ~ c ) c ∪ C ~ , \widetilde{A} \cup \widetilde{B} , \widetilde{A} \cap \widetilde{B}, (\widetilde{A} \cup \widetilde{B})^c, (\widetilde{A} \cap \widetilde{B})^c, (\widetilde{A} \cup \widetilde{B})^c \cap \widetilde{C}, (\widetilde{A} \cup \widetilde{B})^c \cup \widetilde{C}, (\widetilde{A} \cap \widetilde{B}^c)^c \cup \widetilde{C}, A ∪B ,A ∩B ,(A ∪B )c,(A ∩B )c,(A ∪B )c∩C ,(A ∪B )c∪C ,(A ∩B c)c∪C ,
【解】:
-
A ~ ∪ B ~ = ∑ i = 1 6 A ~ ( x i ) x i ∪ ∑ i = 1 6 B ~ ( x i ) x i = 0.9 x 1 + 0.7 x 2 + 0.8 x 3 + 0.9 x 4 + 0.7 x 5 + 0.1 x 6 \widetilde{A} \cup \widetilde{B}= \sum _{i=1} ^6 \frac {\widetilde{A}(x_i)} {x_i} \cup \sum _{i=1} ^6 \frac {\widetilde{B}(x_i)} {x_i} = \frac {0.9} {x_1} +\frac {0.7} {x_2} +\frac {0.8} {x_3} +\frac {0.9} {x_4} +\frac {0.7} {x_5} +\frac {0.1} {x_6} A ∪B =i=1∑6xiA (xi)∪i=1∑6xiB (xi)=x10.9+x20.7+x30.8+x40.9+x50.7+x60.1
-
A ~ ∩ B ~ = ∑ i = 1 6 A ~ ( x i ) x i ∩ ∑ i = 1 6 B ~ ( x i ) x i = 0.1 x 1 + 0.6 x 2 + 0.5 x 3 + 0.2 x 4 + 0.1 x 5 + 0 x 6 \widetilde{A} \cap \widetilde{B} = \sum _{i=1} ^6 \frac {\widetilde{A}(x_i)} {x_i} \cap \sum _{i=1} ^6 \frac {\widetilde{B}(x_i)} {x_i}= \frac {0.1} {x_1} +\frac {0.6} {x_2} +\frac {0.5} {x_3} +\frac {0.2} {x_4} +\frac {0.1} {x_5} +\frac {0} {x_6} A ∩B =i=1∑6xiA (xi)∩i=1∑6xiB (xi)=x10.1+x20.6+x30.5+x40.2+x50.1+x60
-
( A ~ ∪ B ~ ) c = A ~ c ∩ B ~ c = ( 0.9 x 1 + 0.4 x 2 + 0.2 x 3 + 0.1 x 4 + 0.3 x 5 + 0.9 x 6 ) ∩ ( 0.1 x 1 + 0.3 x 2 + 0.5 x 3 + 0.8 x 4 + 0.9 x 5 + 1 x 6 ) = 0.1 x 1 + 0.3 x 2 + 0.2 x 3 + 0.1 x 4 + 0.3 x 5 + 0.9 x 6 (\widetilde{A} \cup \widetilde{B})^c= \widetilde{A}^c \cap \widetilde{B}^c =(\frac {0.9} {x_1} +\frac {0.4} {x_2} +\frac {0.2} {x_3} +\frac {0.1} {x_4} +\frac {0.3} {x_5} +\frac {0.9} {x_6}) \cap (\frac {0.1} {x_1} +\frac {0.3} {x_2} +\frac {0.5} {x_3} +\frac {0.8} {x_4} +\frac {0.9} {x_5} +\frac {1} {x_6}) \\ =\frac {0.1} {x_1} +\frac {0.3} {x_2} +\frac {0.2} {x_3} +\frac {0.1} {x_4} +\frac {0.3} {x_5} +\frac {0.9} {x_6} (A ∪B )c=A c∩B c=(x10.9+x20.4+x30.2+x40.1+x50.3+x60.9)∩(x10.1+x20.3+x30.5+x40.8+x50.9+x61)=x10.1+x20.3+x30.2+x40.1+x50.3+x60.9
-
( A ~ ∩ B ~ ) c = A ~ c ∪ B ~ c = ( 0.9 x 1 + 0.4 x 2 + 0.2 x 3 + 0.1 x 4 + 0.3 x 5 + 0.9 x 6 ) ∪ ( 0.1 x 1 + 0.3 x 2 + 0.5 x 3 + 0.8 x 4 + 0.9 x 5 + 1 x 6 ) = 0.9 x 1 + 0.4 x 2 + 0.5 x 3 + 0.8 x 4 + 0.9 x 5 + 1 x 6 (\widetilde{A} \cap \widetilde{B})^c =\widetilde{A}^c \cup \widetilde{B}^c =(\frac {0.9} {x_1} +\frac {0.4} {x_2} +\frac {0.2} {x_3} +\frac {0.1} {x_4} +\frac {0.3} {x_5} +\frac {0.9} {x_6}) \cup (\frac {0.1} {x_1} +\frac {0.3} {x_2} +\frac {0.5} {x_3} +\frac {0.8} {x_4} +\frac {0.9} {x_5} +\frac {1} {x_6}) \\ =\frac {0.9} {x_1} +\frac {0.4} {x_2} +\frac {0.5} {x_3} +\frac {0.8} {x_4} +\frac {0.9} {x_5} +\frac {1} {x_6} (A ∩B )c=A c∪B c=(x10.9+x20.4+x30.2+x40.1+x50.3+x60.9)∪(x10.1+x20.3+x30.5+x40.8+x50.9+x61)=x10.9+x20.4+x30.5+x40.8+x50.9+x61
-
( A ~ ∪ B ~ ) c ∩ C ~ = ( 0.1 x 1 + 0.3 x 2 + 0.2 x 3 + 0.1 x 4 + 0.3 x 5 + 0.9 x 6 ) ∩ ( 0.6 x 1 + 0.8 x 2 + 1 x 3 + 0.7 x 4 + 0.5 x 5 + 0.3 x 6 ) = 0.1 x 1 + 0.3 x 2 + 0.2 x 3 + 0.1 x 4 + 0.3 x 5 + 0.3 x 6 (\widetilde{A} \cup \widetilde{B})^c \cap \widetilde{C} = (\frac {0.1} {x_1} +\frac {0.3} {x_2} +\frac {0.2} {x_3} +\frac {0.1} {x_4} +\frac {0.3} {x_5} +\frac {0.9} {x_6}) \cap (\frac {0.6} {x_1} +\frac {0.8} {x_2} +\frac {1} {x_3} +\frac {0.7} {x_4} +\frac {0.5} {x_5} +\frac {0.3} {x_6}) \\= \frac {0.1} {x_1} +\frac {0.3} {x_2} +\frac {0.2} {x_3} +\frac {0.1} {x_4} +\frac {0.3} {x_5} +\frac {0.3} {x_6} (A ∪B )c∩C =(x10.1+x20.3+x30.2+x40.1+x50.3+x60.9)∩(x10.6+x20.8+x31+x40.7+x50.5+x60.3)=x10.1+x20.3+x30.2+x40.1+x50.3+x60.3
-
( A ~ ∪ B ~ ) c ∪ C ~ = ( 0.1 x 1 + 0.3 x 2 + 0.2 x 3 + 0.1 x 4 + 0.3 x 5 + 0.9 x 6 ) ∪ ( 0.6 x 1 + 0.8 x 2 + 1 x 3 + 0.7 x 4 + 0.5 x 5 + 0.3 x 6 ) = 0.6 x 1 + 0.8 x 2 + 1 x 3 + 0.7 x 4 + 0.5 x 5 + 0.9 x 6 (\widetilde{A} \cup \widetilde{B})^c \cup \widetilde{C} = (\frac {0.1} {x_1} +\frac {0.3} {x_2} +\frac {0.2} {x_3} +\frac {0.1} {x_4} +\frac {0.3} {x_5} +\frac {0.9} {x_6}) \cup (\frac {0.6} {x_1} +\frac {0.8} {x_2} +\frac {1} {x_3} +\frac {0.7} {x_4} +\frac {0.5} {x_5} +\frac {0.3} {x_6})\\ = \frac {0.6} {x_1} +\frac {0.8} {x_2} +\frac {1} {x_3} +\frac {0.7} {x_4} +\frac {0.5} {x_5} +\frac {0.9} {x_6} (A ∪B )c∪C =(x10.1+x20.3+x30.2+x40.1+x50.3+x60.9)∪(x10.6+x20.8+x31+x40.7+x50.5+x60.3)=x10.6+x20.8+x31+x40.7+x50.5+x60.9
-
( A ~ ∩ B ~ c ) c ∪ C ~ = 0.9 x 1 + 0.8 x 2 + 1 x 3 + 0.7 x 4 + 0.5 x 5 + 0.9 x 6 (\widetilde{A} \cap \widetilde{B}^c)^c \cup \widetilde{C} = \frac {0.9} {x_1} +\frac {0.8} {x_2} +\frac {1} {x_3} +\frac {0.7} {x_4} +\frac {0.5} {x_5} +\frac {0.9} {x_6} (A ∩B c)c∪C =x10.9+x20.8+x31+x40.7+x50.5+x60.9
8.4 已知 A ~ , B ~ ∈ F ( x ) \tilde{A}, \tilde{B} \in F(x) A~,B~∈F(x),
(1)试证 A ~ ∩ B ~ \widetilde{A} \cap \widetilde{B} A ∩B 是包含 A ~ \widetilde{A} A 与 B ~ \widetilde{B} B 的最大公共子集.
证明: ( A ~ ∩ B ~ ) ( x ) = m i n { A ~ ( x ) , B ~ ( x ) } ≤ A ~ ( x ) (\widetilde{A} \cap \widetilde{B} )(x) = min\{\widetilde{A}(x) ,\widetilde{B}(x) \} \le \widetilde{A}(x) (A ∩B )(x)=min{A (x),B (x)}≤A (x) , ( A ~ ∩ B ~ ) ( x ) = m i n { A ~ ( x ) , B ~ ( x ) } ≤ B ~ ( x ) (\widetilde{A} \cap \widetilde{B} )(x) = min\{\widetilde{A}(x) ,\widetilde{B}(x) \} \le \widetilde{B}(x) (A ∩B )(x)=min{A (x),B (x)}≤B (x)
则: A ~ ∩ B ~ ⊆ A ~ \widetilde{A} \cap \widetilde{B} \subseteq \widetilde{A} A ∩B ⊆A , A ~ ∩ B ~ ⊆ B ~ \widetilde{A} \cap \widetilde{B} \subseteq \widetilde{B} A ∩B ⊆B
若 E E E 为任一属于 A ~ \widetilde{A} A 又属于 B ~ \widetilde{B} B 的模糊子集,则有 E ( x ) ≤ A ~ ( x ) E(x) \le \widetilde{A}(x) E(x)≤A (x), E ( x ) ≤ B ~ ( x ) E(x) \le \widetilde{B}(x) E(x)≤B (x)
从而 E ≤ ( A ~ ∩ B ~ ) ( x ) = m i n { A ~ ( x ) , B ~ ( x ) } = A ~ ( x ) ∩ B ~ ( x ) E \le (\widetilde{A} \cap \widetilde{B} )(x) = min\{\widetilde{A}(x) ,\widetilde{B}(x) \}= \widetilde{A}(x) \cap \widetilde{B}(x) E≤(A ∩B )(x)=min{A (x),B (x)}=A (x)∩B (x)
则 E ≤ A ~ ∩ B ~ E \le \widetilde{A} \cap \widetilde{B} E≤A ∩B , 所以 A ~ ∩ B ~ \widetilde{A} \cap \widetilde{B} A ∩B 是包含 A ~ \widetilde{A} A 与 B ~ \widetilde{B} B 的最大公共子集。
(2)试证 A ~ ∪ B ~ \widetilde{A} \cup \widetilde{B} A ∪B 是包含 A ~ \widetilde{A} A 与 B ~ \widetilde{B} B 的最大子集.
证明: ( A ~ ∪ B ~ ) ( x ) = m a x { A ~ ( x ) , B ~ ( x ) } ≥ A ~ ( x ) (\widetilde{A} \cup \widetilde{B} )(x) = max\{\widetilde{A}(x) ,\widetilde{B}(x) \} \ge \widetilde{A}(x) (A ∪B )(x)=max{A (x),B (x)}≥A (x) , ( A ~ ∪ B ~ ) ( x ) = m a x { A ~ ( x ) , B ~ ( x ) } ≥ B ~ ( x ) (\widetilde{A} \cup \widetilde{B} )(x) = max\{\widetilde{A}(x) ,\widetilde{B}(x) \} \ge \widetilde{B}(x) (A ∪B )(x)=max{A (x),B (x)}≥B (x)
则: A ~ ∪ B ~ ⊇ A ~ \widetilde{A} \cup \widetilde{B} \supseteq \widetilde{A} A ∪B ⊇A , A ~ ∪ B ~ ⊇ B ~ \widetilde{A} \cup \widetilde{B} \supseteq \widetilde{B} A ∪B ⊇B
若 E E E 为任一包含 A ~ \widetilde{A} A 又包含 B ~ \widetilde{B} B 的模糊子集,则有 E ( x ) ≥ A ~ ( x ) E(x) \ge \widetilde{A}(x) E(x)≥A (x), E ( x ) ≥ B ~ ( x ) E(x) \ge \widetilde{B}(x) E(x)≥B (x)
从而 E ≥ ( A ~ ∪ B ~ ) ( x ) = m a x { A ~ ( x ) , B ~ ( x ) } = A ~ ( x ) ∪ B ~ ( x ) E \ge (\widetilde{A} \cup \widetilde{B} )(x) = max\{\widetilde{A}(x) ,\widetilde{B}(x) \}= \widetilde{A}(x) \cup \widetilde{B}(x) E≥(A ∪B )(x)=max{A (x),B (x)}=A (x)∪B (x)
则 E ≥ A ~ ∪ B ~ E \ge \widetilde{A} \cup \widetilde{B} E≥A ∪B , 所以 A ~ ∪ B ~ \widetilde{A} \cup \widetilde{B} A ∪B 是包含 A ~ \widetilde{A} A 与 B ~ \widetilde{B} B 的最大子集。
8.5 设 X = [ 0 , 1 ] , A ~ ( x ) = x , A c ~ ( x ) = 1 − x X=[0,1],\tilde{A}(x)=x,\tilde{A^c}(x)=1-x X=[0,1],A~(x)=x,Ac~(x)=1−x,试证不满足互补率.
解 : 由题设可得
(
A
~
∪
A
~
c
)
(
x
)
=
{
1
−
x
,
当
x
⩽
1
2
,
x
,
当
>
1
2
.
(
A
~
∩
A
~
c
)
(
x
)
=
{
x
,
当
x
⩽
1
2
,
1
−
x
,
当
>
1
2
,
\begin{array}{l} \left(\widetilde{A} \cup \tilde{A}^{c}\right)(x)=\left\{\begin{array}{cc} 1-x, & \text { 当 } x \leqslant \frac{1}{2}, \\ x, & \text { 当 }>\frac{1}{2} . \end{array}\right. \\ \left(\widetilde{A} \cap \tilde{A}^{c}\right)(x)=\left\{\begin{array}{cc} x, & \text { 当 } x \leqslant \frac{1}{2}, \\ 1-x, & \text { 当 }>\frac{1}{2}, \end{array}\right. \end{array}
(A
∪A~c)(x)={1−x,x, 当 x⩽21, 当 >21.(A
∩A~c)(x)={x,1−x, 当 x⩽21, 当 >21,
可得
A
~
∪
A
~
c
≠
X
;
A
~
∩
A
~
c
≠
∅
\widetilde{A} \cup \tilde{A}^{c} \neq X ;\qquad \widetilde{A} \cap \tilde{A}^{c} \neq \varnothing
A
∪A~c=X;A
∩A~c=∅ 则不满足互补率。
8.6 欲证 ( A ~ ∪ B ~ ) ∩ C ~ = ( A ~ ∩ C ~ ) ∪ ( B ~ ∩ C ~ ) (\tilde{A} \cup \tilde{B}) \cap \tilde{C}=(\tilde{A} \cap \tilde{C}) \cup(\tilde{B} \cap \tilde{C}) (A~∪B~)∩C~=(A~∩C~)∪(B~∩C~)
证明 : => A ~ ∪ B ~ ⊃ A ~ \tilde{A} \cup \tilde{B} \supset \tilde{A} A~∪B~⊃A~
( A ~ ∪ B ~ ) ∩ C ~ ⊃ A ~ ∪ C ~ (\tilde{A} \cup \tilde{B})\cap \tilde{C} \supset \tilde{A} \cup \tilde{C} (A~∪B~)∩C~⊃A~∪C~
A ~ ∪ B ~ ⊃ B ~ \tilde{A} \cup \tilde{B} \supset \tilde{B} A~∪B~⊃B~
( A ~ ∪ B ~ ) ∩ C ~ ⊃ B ~ ∪ C ~ (\tilde{A} \cup \tilde{B})\cap \tilde{C} \supset \tilde{B} \cup \tilde{C} (A~∪B~)∩C~⊃B~∪C~
∴ ( A ~ ∪ B ~ ) ∩ C ~ ⊃ ( A ~ ∩ C ~ ) ∪ ( B ~ ∩ C ~ ) \therefore (\tilde{A} \cup \tilde{B})\cap \tilde{C} \supset (\tilde{A} \cap \tilde{C}) \cup (\tilde{B} \cap \tilde{C}) ∴(A~∪B~)∩C~⊃(A~∩C~)∪(B~∩C~)
<= X ∈ ( A ~ ∪ B ~ ) ∩ C ~ X \in (\tilde{A} \cup \tilde{B})\cap \tilde{C} X∈(A~∪B~)∩C~ X ∈ ( A ~ ∪ B ~ ) X \in (\tilde{A} \cup \tilde{B}) X∈(A~∪B~) 且 x ∈ C x \in C x∈C
X ∈ A ~ X \in \tilde{A} X∈A~ 或 X ∈ B ~ X \in \tilde{B} X∈B~ 且 X ∈ C ~ X \in \tilde{C} X∈C~
当 x ∈ A ~ x \in \tilde{A} x∈A~ 时, x ∈ ( A ~ ∩ C ~ ) x \in (\tilde{A} \cap \tilde{C}) x∈(A~∩C~) , 当 x ∈ B ~ x \in \tilde{B} x∈B~ 时, x ∈ ( B ~ ∩ C ~ ) x \in (\tilde{B} \cap \tilde{C}) x∈(B~∩C~) ,
∴ X ∈ ( A ~ ∩ C ~ ) ∪ ( B ~ ∩ C ~ ) \therefore X \in (\tilde{A} \cap \tilde{C}) \cup (\tilde{B} \cap \tilde{C}) ∴X∈(A~∩C~)∪(B~∩C~)
由上面可得: ( A ~ ∪ B ~ ) ∩ C ~ = ( A ~ ∩ C ~ ) ∪ ( B ~ ∩ C ~ ) (\tilde{A} \cup \tilde{B}) \cap \tilde{C}=(\tilde{A} \cap \tilde{C}) \cup(\tilde{B} \cap \tilde{C}) (A~∪B~)∩C~=(A~∩C~)∪(B~∩C~)
8.7 设
X
=
{
1
,
2
,
3
,
4
,
5
}
X=\{1,2,3,4,5\}
X={1,2,3,4,5},
A
~
=
0.2
x
1
+
0.7
x
2
+
1
x
3
+
0.8
x
4
+
0.3
x
5
,
B
~
=
0.7
x
1
+
1
x
2
+
0.1
x
3
+
1
x
4
+
0.6
x
5
,
\tilde{A} = \frac {0.2} {x_1} + \frac {0.7} {x_2} +\frac {1} {x_3}+\frac {0.8} {x_4}+\frac {0.3} {x_5}, \\ \tilde{B} = \frac {0.7} {x_1} + \frac {1} {x_2} +\frac {0.1} {x_3}+\frac {1} {x_4}+\frac {0.6} {x_5},
A~=x10.2+x20.7+x31+x40.8+x50.3,B~=x10.7+x21+x30.1+x41+x50.6,
求
A
~
∧
B
~
,
A
~
∨
B
~
,
A
~
⊕
B
~
,
A
~
⊙
B
~
\tilde{A} \wedge \tilde{B}, \tilde{A} \vee \tilde{B},\tilde{A} \oplus \tilde{B},\tilde{A} \odot \tilde{B}
A~∧B~,A~∨B~,A~⊕B~,A~⊙B~
解:
A
~
∧
B
~
=
0.14
x
1
+
0.7
x
2
+
0.1
x
3
+
0.8
x
4
+
0.18
x
5
A
~
∨
B
~
=
0.7
x
1
+
1
x
2
+
1
x
3
+
1
x
4
+
0.6
x
5
A
~
⊕
B
~
=
0.9
x
1
+
1
x
2
+
1
x
3
+
1
x
4
+
0.9
x
5
A
~
⊙
B
~
=
0
x
1
+
0.7
x
2
+
0.1
x
3
+
0.8
x
4
+
0
x
5
=
0.7
x
2
+
0.1
x
3
+
0.8
x
4
\tilde{A} \wedge \tilde{B} = \frac {0.14} {x_1} + \frac {0.7} {x_2} +\frac {0.1} {x_3}+\frac {0.8} {x_4}+\frac {0.18} {x_5} \\ \tilde{A} \vee \tilde{B}=\frac {0.7} {x_1} + \frac {1} {x_2} +\frac {1} {x_3}+\frac {1} {x_4}+\frac {0.6} {x_5} \\ \tilde{A} \oplus \tilde{B}= \frac {0.9} {x_1} + \frac {1} {x_2} +\frac {1} {x_3}+\frac {1} {x_4}+\frac {0.9} {x_5} \\ \tilde{A} \odot \tilde{B} = \frac {0} {x_1} + \frac {0.7} {x_2} +\frac {0.1} {x_3}+\frac {0.8} {x_4}+\frac {0} {x_5}=\frac {0.7} {x_2}+\frac {0.1} {x_3}+\frac {0.8} {x_4}
A~∧B~=x10.14+x20.7+x30.1+x40.8+x50.18A~∨B~=x10.7+x21+x31+x41+x50.6A~⊕B~=x10.9+x21+x31+x41+x50.9A~⊙B~=x10+x20.7+x30.1+x40.8+x50=x20.7+x30.1+x40.8