Quantcast
Channel: Active questions tagged real-analysis - Mathematics Stack Exchange
Viewing all articles
Browse latest Browse all 8490

Tips on showing collections are $\sigma$-algebras and determining a $\sigma$-algebra generated by a collection

$
0
0

So I am currently enrolled in a measure theory class and as such, a number of homework questions I've come across so far have been along the lines of 'show that some collection is a $\sigma$-algebra on some set' or 'determine the $\sigma$-algebra generated by some collection.' I seem to have no trouble showing this when the collection is something simple, like $\{ \emptyset,\{a\},\{b\}, \{a,b\} \}$ for $a,b\in X=\{a,b \}\subset\mathbb{R}$. However, things apparently are more difficult for me to grasp when the collection is very large.

For instance, on my previous homework, I was asked to show that $\left\{\bigcup_{n\in K} [n,n+1)\Bigg{|}K\subseteq\textbf{Z} \right\}$is a $\sigma$-algebra on $\mathbb{R}$. My approach was as follows (and this is exactly what I handed in):

First, note that\begin{align*} \textbf{R}\setminus\bigcup_{n\in K}[n,n+1) &= \textbf{R}_{<\inf K}\cup \textbf{R}_{\geq(\sup K)+1} \\ &= \cdots\cup[\inf K -2,\inf K-1)\cup[\inf K-1,\inf K)\cup[\sup K+1,\sup K+2)\cup[\sup K+2,\sup K+3)\cup\cdots\end{align*}This shows closure under complementation. Now, let $K_1,K_2,\cdots$ be a sequence of subsets of \textbf{Z}. Then\begin{equation*} \bigcup_{i=1}^\infty \bigcup_{n\in K_i}[n,n+1) = \bigcup_{n\in\bigcup_{i=1}^\infty K_i}[n,n+1)\end{equation*}But $\bigcup_{i=1}^\infty K_i\subset\textbf{Z}$, so $\bigcup_{n\in\bigcup_{i=1}^\infty K_i}[n,n+1)\in \left\{\bigcup_{n\in K} [n,n+1)|K\subseteq\textbf{Z} \right\}$. This shows closure under countable unions. Finally, since $\emptyset\subset\textbf{Z}$, $\bigcup_{n\in\emptyset}[n,n+1)=\emptyset$. So the collection is indeed a $\sigma$-algebra.

I lost a number of points on this last homework, and I presume it's due to a question like this.

Similarly, another question I had was to determine the $\sigma$-algebra on R generated by(a) $A:=\{(r,r+1)|r\in\textbf{Q} \}$ and (b) $B:=\{[r,\infty)|r\in\textbf{Q} \}$. I answered as follows:

Any $\sigma$-algebra containing $A$ must contain $\emptyset, \bigcup_{k=1}^\infty (r_k,r_k+1)$, and $(r,r+1)^\complement=\textbf{R}\setminus(r,r+1)$. For $r_1<r_1+1<r_2\in\textbf{Q}$, we have that $(r_1,r_1+1)\cup(r_2,r_2+1)=(r_1,r_2+1)$, which is just another open rational interval. We could do this union for countably many open rational intervals and still get another open rational interval. Now,\begin{equation*} (r,r+1)^\complement = \textbf{R}\setminus(r,r+1)=\left\{\bigcup_{i=1}^\infty (r_i,r_i+1)\Bigg{|}r_i\neq r, r_i+1\neq r+1 \right\} = \text{ another open rational interval.}\end{equation*}Finally, $\bigcap_{i=1}^\infty (r_i,r_i+1)=\emptyset$. Since in this $\sigma$-algebra rational open intervals are closed under the $\sigma$-algebra operations and $\overline{\textbf{Q}}=\textbf{R}$, the smallest $\sigma$-algebra containing $A$ is the open sets of $\mathbb{R}$, that is the Borel $\sigma$-algebra.

For part (b), note that\begin{equation*} \bigcup_{k=1}^\infty [r_k,\infty) = \bigcup_{k=1}^\infty\bigcap_{j=1}^\infty \left(r_k+\frac{1}{j},\infty \right) = \bigcap_{j=1}^\infty\left(\bigcup_{k=1}^\infty \left(r_k+\frac{1}{j},\infty \right) \right)\end{equation*}Similarly,\begin{equation*} [r,\infty)^\complement = \bigcup_{j=1}^\infty \textbf{R}\setminus\left(r_k+\frac{1}{j},\infty \right)=\lim_{N\to\infty}\bigcup_{j=1}^N \textbf{R}\setminus\left(r+\frac{1}{j},\infty \right)=\textbf{R}\setminus(r,\infty)=(-\infty,r)\end{equation*}which can be expressed as a countable union of rational open intervals which is another rational open interval. Finally, $(-\infty,r)\cap[r,\infty)=\emptyset$. Since the intervals in $B$ can be expressed as countable intersections of open rational intervals which we have just demonstrated are closed under countable unions and complementation, we conclude that the smallest $\sigma$-algebra containing $B$ is again the Borel $\sigma$-algebra.

I feel like my approaches are all wrong. If anyone could give some tips on how best to approach problems like this, I'd appreciate it.


Viewing all articles
Browse latest Browse all 8490

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>