概率论基础-严士健 第二版 习题与补充1.6答案

概率论基础-严士健 第二版
习题与补充1.6答案

1.不独立。
设有一均匀正八面体,其第1,2,3,4面染红色,第1,2,3,5面染白色,第1,6,7,8面染黑色,现在以A,B,C分别表示投一次正八面体出现红、白、黑的事件,则
P ( A ) = P ( B ) = P ( C ) = 1 2 . P(A) = P(B) = P(C) = \frac{1}{2}. P(A)=P(B)=P(C)=21.
P ( A B C ) = 1 8 = P ( A ) P ( B ) P ( C ) . P(ABC) = \frac{1}{8} = P(A)P(B)P(C). P(ABC)=81=P(A)P(B)P(C).
但是,
P ( A B ) = 3 8 ≠ 1 4 = P ( A ) P ( B ) P(AB) = \frac{3}{8} \neq \frac{1}{4} = P(A)P(B) P(AB)=83=41=P(A)P(B)
故A、B、C不两两独立。
2. P ( A n i . o . ) = P ( ∪ n = 1 ∞ ∩ k ≥ n A k ) = lim ⁡ n → ∞ P ( ∩ k ≥ n A k ) ≤ lim ⁡ n → ∞ ∑ k ≥ n P ( A k ) = 0. P(A_n i.o.) = P(\cup_{n=1}^{\infty}\cap_{k\geq n} A_k) = \lim_{n\to \infty} P(\cap_{k\geq n} A_k) \leq \lim_{n\to \infty} \sum_{k\geq n}P( A_k) = 0. P(Ani.o.)=P(n=1knAk)=limnP(knAk)limnknP(Ak)=0.
P ( A n i . o . ) = 1 − lim ⁡ n → ∞ lim ⁡ m → ∞ Π k = n m P ( A k c ) = 1 − lim ⁡ n → ∞ lim ⁡ m → ∞ Π k = n m ( 1 − P ( A k ) ) ≥ 1 − lim ⁡ n → ∞ lim ⁡ m → ∞ Π k = n m e − P ( A k ) = 1 − lim ⁡ n → ∞ lim ⁡ m → ∞ e − ∑ k = n m P ( A k ) = 1. P(A_n i.o.) = 1 - \lim_{n\to \infty} \lim_{m\to \infty} \Pi_{k=n}^m P(A_k^c) = 1 - \lim_{n\to \infty} \lim_{m\to \infty} \Pi_{k=n}^m (1 - P(A_k)) \geq 1 - \lim_{n\to \infty} \lim_{m\to \infty} \Pi_{k=n}^m e^{-P(A_k)} = 1 - \lim_{n\to \infty} \lim_{m\to \infty} e^{-\sum_{k=n}^mP(A_k)} = 1. P(Ani.o.)=1limnlimmΠk=nmP(Akc)=1limnlimmΠk=nm(1P(Ak))1limnlimmΠk=nmeP(Ak)=1limnlimmek=nmP(Ak)=1.
3.i) Π k = 1 n ( 1 − p k ) \Pi_{k=1}^n (1 - p_k) Πk=1n(1pk)
ii)1 - Π k = 1 n ( 1 − p k ) \Pi_{k=1}^n (1 - p_k) Πk=1n(1pk)
iii) ∑ k = 1 n p k Π j ≠ k ( 1 − p j ) \sum_{k=1}^np_k \Pi_{j\neq k} (1-p_j) k=1npkΠj=k(1pj)
4.由定义3知 C t , t ∈ T \mathscr{C}_t, t \in T Ct,tT独立的充分必要条件是对 T T T的任意有限子集 { t 1 , … , t n } , n ≥ 1 , C t k , k = 1 , … , n \{t_1, \dots, t_n\}, n \geq 1, \mathscr{C}_{t_k}, k = 1, \dots, n {t1,,tn},n1,Ctk,k=1,,n独立,因而只需对 T = { 1 , … , n } T = \{1, \dots, n\} T={1,,n}的情形证明。
对任意 A j ∈ C j , j = 1 , … , n A_j \in \mathscr{C}_j, j = 1, \dots, n AjCj,j=1,,n,任取 1 ≤ k 1 < k 1 < ⋯ < k m ≤ n 1 \leq k_1 < k_1 < \cdots < k_m \leq n 1k1<k1<<kmn
i)设事件A的概念为0,事件B 的概念为1。则
P ( ∩ j = 1 m A k j ∩ A ∩ B ) = 0 = Π j = 1 m P ( A k j ) P ( A ) P ( B ) ; P(\cap_{j=1}^m A_{k_j} \cap A \cap B) = 0 = \Pi_{j=1}^m P(A_{k_j})P(A)P(B); P(j=1mAkjAB)=0=Πj=1mP(Akj)P(A)P(B);
P ( ∩ j = 1 m A k j = Π j = 1 m P ( A k j ) ; P(\cap_{j=1}^m A_{k_j}= \Pi_{j=1}^m P(A_{k_j}); P(j=1mAkj=Πj=1mP(Akj);
P ( ∩ j = 1 m A k j ∩ A ) = 0 = Π j = 1 m P ( A k j ) P ( A ) ; P(\cap_{j=1}^m A_{k_j} \cap A) = 0 = \Pi_{j=1}^m P(A_{k_j})P(A); P(j=1mAkjA)=0=Πj=1mP(Akj)P(A);
P ( ∩ j = 1 m A k j ∩ B ) = P ( ∩ j = 1 m A k j ) = Π j = 1 m P ( A k j ) P ( B ) . P(\cap_{j=1}^m A_{k_j}\cap B) = P(\cap_{j=1}^m A_{k_j}) = \Pi_{j=1}^m P(A_{k_j})P(B). P(j=1mAkjB)=P(j=1mAkj)=Πj=1mP(Akj)P(B).
ii) 设 A ⊃ B A \supset B AB A ∈ C j A \in \mathscr{C}_j ACj,则 P ( ∩ j = 1 m A k j ∩ ( A − B ) ) = P ( ( ∩ j = 1 m A k j ∩ A ) − ( ∩ j = 1 m A k j ∩ B ) ) = P ( ∩ j = 1 m A k j ∩ A ) − P ( ∩ j = 1 m A k j ∩ B ) = Π j = 1 m P ( A k j ) P ( A ) − Π j = 1 m P ( A k j ) P ( B ) = Π j = 1 m P ( A k j ) ( P ( A ) − P ( B ) ) = Π j = 1 m P ( A k j ) P ( A − B ) . P(\cap_{j=1}^m A_{k_j} \cap (A - B)) = P((\cap_{j=1}^m A_{k_j} \cap A ) - (\cap_{j=1}^m A_{k_j} \cap B)) = P(\cap_{j=1}^m A_{k_j} \cap A ) - P(\cap_{j=1}^m A_{k_j} \cap B) = \Pi_{j=1}^mP(A_{k_j})P(A) -\Pi_{j=1}^mP(A_{k_j})P(B) =\Pi_{j=1}^mP(A_{k_j})(P(A) - P(B)) =\Pi_{j=1}^mP(A_{k_j})P(A - B). P(j=1mAkj(AB))=P((j=1mAkjA)(j=1mAkjB))=P(j=1mAkjA)P(j=1mAkjB)=Πj=1mP(Akj)P(A)Πj=1mP(Akj)P(B)=Πj=1mP(Akj)(P(A)P(B))=Πj=1mP(Akj)P(AB).
iii)及iv)考虑由 C \mathscr{C} C所生成的 σ \sigma σ代数,再由独立性的证明立马得证。
5. G k = { ∩ j = i m i A t j , A t j ∈ C t j , t j ∈ T k , m ≥ 1 } \mathscr{G_k} = \{\cap_{j=i}^{m_{i}} A_{t_j}, A_{t_j} \in \mathscr{C}_{t_j}, t_j \in T_k, m \geq 1\} Gk={j=imiAtj,AtjCtj,tjTk,m1},只需证 G k \mathscr{G_k} Gk相互独立且每一个 G k \mathscr{G_k} Gk Π \Pi Π系, k = 1 , … , n k = 1, \dots, n k=1,,n
i)取任一有限子集 { i 1 , … , i n } \{i_1, \dots, i_n\} {i1,,in},由于事件类 C t , t ∈ T \mathscr{C}_t, t \in T Ct,tT独立,从而 C t i 1 , … , C t i n \mathscr{C}_{t_{i_1}}, \dots, \mathscr{C}_{t_{i_n}} Cti1,,Ctin独立。则 P ( ∩ j = i 1 m i 1 A t j ∩ … , ∩ j = i n m i n A t j ) = P ( ∩ j = i 1 m i 1 A t j ) … P ( ∩ j = i n m i n A t j ) P(\cap_{j=i_1}^{m_{i_1}} A_{t_j} \cap \dots, \cap_{j=i_n}^{m_{i_n}}A_{t_j}) = P(\cap_{j=i_1}^{m_{i_1}} A_{t_j}) \dots P(\cap_{j=i_n}^{m_{i_n}}A_{t_j}) P(j=i1mi1Atj,j=inminAtj)=P(j=i1mi1Atj)P(j=inminAtj).
ii)取 A = ∩ j = 1 m A t j , B = ∩ i = 1 h B t i , A t j ∈ C t j , B t i ∈ C t i , t j , t i ∈ T k A = \cap_{j=1}^m A_{t_j}, B = \cap_{i=1}^h B_{t_i}, A_{t_j} \in \mathscr{C}_{t_j}, B_{t_i} \in \mathscr{C}_{t_i}, t_j, t_i \in T_k A=j=1mAtj,B=i=1hBti,AtjCtj,BtiCti,tj,tiTk,则
A ∩ B = ∩ j = 1 m A t j ∩ ∩ i = 1 h B t i = ∩ j = 1 m ∩ i = 1 h A t j ∩ B t i = ∩ j = 1 max ⁡ { m , h } A t j ∩ B t j A \cap B = \cap_{j=1}^m A_{t_j} \cap \cap_{i=1}^h B_{t_i} = \cap_{j=1}^m \cap_{i=1}^h A_{t_j} \cap B_{t_i} =\cap_{j=1}^{\max\{m,h\}}A_{t_j} \cap B_{t_j} AB=j=1mAtji=1hBti=j=1mi=1hAtjBti=j=1max{m,h}AtjBtj得到 Π \Pi Π系。

你可能感兴趣的:(概率论基础,严士健,习题与补充)