6

I'm beginning to study about stochastic processes, and currently focusing on stopping times and hitting times. The textbook I'm using is "Stochastic Integration Theory" by Medvegyev (and Karatzas & Shreve as a second reference), and in some of the theorems the following measurable projection theorem is used.

If the space $(\Omega,\mathcal{A},\mathbb{P})$ is complete and $$U \in \mathcal{B}(\mathbb{R}^n) \otimes \mathcal{A},$$ then $$\text{proj}_\Omega (U) := \{x: \exists t\text{ such that }(t,x) \in U\} \in \mathcal{A}.$$

On the authors homepage there is a note containing a proof as well as many definitions such as Suslin (also called analytic) sets and auxiliary lemmas, however I find the material to be lacking in rigor and it is missing some assumptions. Therefore I am looking for a textbook in which the measurable projection is covered in detail. I've looked at the textbooks by Kechris, and Srivastava without finding what I was looking for.

Olorun
  • 1,621
  • Dellacherie and Meyer, Probabilities and Potential, Chapter III. – zhoraster Jan 08 '16 at 07:24
  • I didn't find this in Chapter III, but should certainly be inside this book. All such facts are there. – zhoraster Jan 08 '16 at 07:40
  • Yeah I just looked at Chapter III, but to be honest, the notation and organization is a little messy so it's hard to find the actual theorem. – Olorun Jan 08 '16 at 07:54
  • @Olorun Agreed with not best place... Anyways,,, suggestion. Take it or leave it. I had got the same helpful hint on Math.SE and Chemistry.SE both, so thought of passing it on... – Shailesh Jan 13 '16 at 04:40

1 Answers1

4

Theorem 13 in Chapter III of the first volume of Dellacherie & Meyer (cited by @zhoraster; see the foot of page 43 in the English translation) tells you that the projection onto $\Omega$ of $U$ is $\mathcal A$-analytic. As such, this projection is $\mathcal A$-measurable, because $(\Omega,\mathcal A,\Bbb P)$ is complete; see no. III-33 at the top of p. 58 of D. & M.

John Dawkins
  • 29,845
  • 1
  • 23
  • 39