练习

Give an example of a martingale with and for . This example shows that it is not enough to have in Theorem 4.3.1.

证明

。若 ,则 ;若 ,则 . 可以由 Borel-Cantelli 引理得到,而 是因为 然后用 Borel-Cantelli 引理。 是因为 用 Borel-Cantelli 引理,把这部分排除掉剩余部分满足 a.s..

练习

(Assumes familiarity with finite state Markov chains.) Fine tune the example for the previous problem so that and , where is your favorite number in , i.e., you are asked to do this for one value of that you may choose. This example shows that a martingale can converge in distribution without converging a.s. (or in probability).

证明

TODO

练习

Let and be positive integrable and adapted to . Suppose , with a.s. . Prove that converges a.s. to a finite limit. Hint: Let , and stop your supermartingale at time .

证明

首先构造上鞅 . 由

知道 是上鞅。对 ,令停时

那么 ,那么

是上鞅,并且 a.s. 收敛,那么 上 a.e. 收敛。令 得到 上 a.s. 收敛,那么 a.s. 收敛。

练习

Let . Use the Borel-Cantelli lemmas to show that

证明

为互相独立的事件列,且 ,那么由 Borel-Cantelli 引理知道

这等价于

练习

Show

implies

证明

,令

那么由 和上一题结论知道

从而

练习

Check by direct computation that the in Example 4.3.7 is a martingale. Show that if we drop the condition and set when , then .

证明

对第一问,只要证明 ,也即

由于 可以由有限个元素生成,无妨设 . 那么

从而得证。

对第二问,仍然考虑上式。 处等号应改成 ,因为等式右侧只包含所有 的区间对应的项。

练习

Apply Theorem 4.3.5 to Example 4.3.7 to get a "probabilistic" proof of the Radon-Nikodym theorem. To be precise, suppose is countably generated (i.e., there is a sequence of sets so that ) and show that if and are -finite measures and , then there is a function so that . Before you object to this as circular reasoning (the Radon - Nikodym theorem was used to define conditional expectation!), observe that the conditional expectations that are needed for Example 4.3.7 have elementary definitions.

证明

-有限性,可以不妨设 是概率测度。令 分别为 上的限制。那么有 成立,再令 ,那么由定理 4.3.5 知道

由于 知道 的 Lebesgue 分解存在且奇异部分为 ,从而

从例 4.3.7 也可以得出同样的结论。

For the next three exercises, suppose are concentrated on and have , .

练习

  1. Use Theorem 4.3.8 to find a necessary and sufficient condition for .
  2. Suppose that . Show that in this case the condition is simply .

证明

对于第一问,首先有

然后直接代入定理得到等价条件

第二问 TODO

练习

Show that if and in the previous exercise then . This shows that the condition is not sufficient for in general.

证明

直接由下一问可得。

练习

Suppose . Show that is sufficient for in general.

证明

不妨设 来去除绝对值。由 知道

其中 表示敛散性相同。同理可以得到

那么两式相加直接得到

练习

Show that if then it is and hence

证明

,那么

那么 . 之后是容易证明的。

练习

Let be a branching process with offspring distribution , defined in part d of Section 4.3, and let . Suppose has . Show that is a martingale and use this to conclude .

证明

先证鞅性,有

对于第二问,由 是非负鞅可以设 ,下分析 的分布。考虑集合 ,此时 ,而在 上,一定有 从而 ,且 ,从而 . 那么只要证明 . 这是容易的,由控制收敛定理得到 .

练习

Galton and Watson who invented the process that bears their names were interested in the survival of family names. Suppose each family has exactly 3 children but coin flips determine their sex. In the 1800s, only male children kept the family name so following the male offspring leads to a branching process with , , , . Compute the probability that the family name will die out when .

证明

直接解方程

解出 .