Let (X,M) and (Y,N) be measurable spaces, and let f:X×Y→[−∞,∞] be some measurable function. Then we may define ˉf:Y→[−∞,∞] by ˉf(y)=inf but is this function \mathscr N-measurable? I suspect the answer is yes, though I don't see why.
Edit: Note that the statement that the projection map \pi:X\times Y\to Y takes measurable sets to measurable sets is false.
Answer
The statement is in general false: Let X= Y= \mathbb{R} and \mathcal{M} the Lebesgue-\sigma-algebra and \mathcal{N} the Borel-\sigma-algebra. Now take any set E \subset \mathbb{R} which is Lebesgue-measurable, but not Borel.
Define f(x,y) = \begin{cases} 0 & \text{if } x=y \\ 1 & \text{else} \end{cases}.
Of course, this map is measurable according to \mathcal{M} \otimes \mathcal{N}, because \{(x,y) \in \mathbb{R} : x=y\} is closed and thus a Borel-set. On the other side \inf_{x \in E} f(x,y) = 1-1_{E}(y)
is not Borel-measurable.
No comments:
Post a Comment