I picked up this book called "Lebesgue Integration on Euclidean Space" but I cannot find a solutions manual for it anywhere. Problem 2 talks about limsup and liminf.

It goes like this: for a sequence of sets $A_1,A_2,...$, prove that

$$\cap_{j=1}^{\infty} \left( \cup_{k=j}^{\infty} A_k \right) = \{ x| x \in A_k \text{for infinitely many} \ \ k \}$$

My attempt at it:

If we look at the union of sets inside the brackets then we will have:

$$A_j \cup A_{j+1} \cup ...$$

and so we can rewrite LHS:

$$\cap_{j=1}^{\infty} \left( A_j \cup A_{j+1} \cup ... \right) = (A_1 \cup A_{2} \cup ...) \cap (A_2 \cup A_3 \cup ...) \cap ...$$

And so naturally due to intersection we will only have $x \in A_k$ for infinitely large $k$ in our final set, however, that is not what the problem asks to prove. It asks to prove that the final set will be composed of $x \in A_k$ for infinitely many $k$. So I do not think I have proved anything.

  • $\begingroup$ Does a sequence of sets in this case mean $A_1 \subset A_2 \subset A_3 ...$ ? What if one set from this sequence is $\emptyset$ $\endgroup$ – Naz Jan 29 '17 at 20:28
  • $\begingroup$ Is the index i used in the intersection supposed to be j? I don't see i used anywhere else $\endgroup$ – pwerth Jan 29 '17 at 20:30

Suppose $x \in A_i$ for infinitely many $i$. Then $x \in \bigcup_{k = j}^\infty A_k$ since we find an $i \geq k$ because there are infinitely many ones. So especially $x \in \bigcap_{j = 1}^\infty \bigcup_{k = j}^\infty A_k$. Conversly suppose $x \in \bigcap_{j = 1}^\infty \bigcup_{k = j}^\infty A_k$. Then $x \in \bigcup_{k = j}^\infty A_k$ for any $j \in \mathbb{N}$. Which means that $x \in A_k$ for some $k \leq j$ for all $j \in \mathbb{N}$. Hence $x \in A_k$ for infinitely many $k$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.