What is the Relationship Between Liminf and Limsup in Subset Sequences?

  • Thread starter Thread starter SpaceWalrus
  • Start date Start date
AI Thread Summary
The relationship between liminf and limsup in sequences of subsets is clarified, showing that liminf A_n is a subset of limsup A_n. An example sequence is provided where A_n alternates between the empty set and the whole space, demonstrating that both liminf and limsup equal the whole space. The confusion around definitions is addressed, with limsup defined as points in infinitely many sets and liminf as points in all but finitely many sets. A proof approach is suggested, emphasizing that as n approaches infinity, the inclusion relationship holds. Overall, the discussion resolves the initial confusion about the definitions and their implications.
SpaceWalrus
Messages
18
Reaction score
0
I have been told that given a sequence of subsets, liminf A_n\subseteq limsup A_n. I have spent some time trying to cook up such a sequence, but I can only ever show they are equal. Can someone point me in the right direction?
 
Physics news on Phys.org
If there is no restriction on An, then you could set up a sequence:
An = empty set for even n and An = whole space for odd n.
 
mathman said:
If there is no restriction on An, then you could set up a sequence:
An = empty set for even n and An = whole space for odd n.

I don't know if that one works. Note that

liminf A_n = A_1 \cup [A_1 \cap A_2] \cup [A_1 \cap A_2 \cap A_3] \dots = X \cup [X \cap \emptyset] \cup [X \cap \emptyset \cap X] \dots = X \cup \emptyset \cup \emptyset \dots = X,

while

limsupA_n = A_1 \cap [A_1 \cup A_2] \cap [A_1 \cup A_2 \cup A_3] \dots = X\cap [X \cup \emptyset] \cap [X \cup \emptyset \cup X] \dots = X \cap X \cap X \dots = X.

Hence they are equal.
 
I am a little confused by your definitions. My understanding is that limsup consists of all points which are in an infinite number of sets of the sequence, which would be the whole space. Liminf according to my understanding are those points which belong to all but a finite number of sets in the sequence, which would be none.

To use your expression I think you need to start with An (not A1) and then let n become infinite.
 
Last edited:
Yeah, I had the definition wrong. UGH! Thanks!
 
How can I prove that liminf of a sequence of events An is a subset of lim sup of An given n goes to infinity? I'm inexperienced in constructing proofs, so please do not be strict.
 
One way of getting there is to look at
supn = An∪An+1∪...
infn = An∩An+1∩...
supn contains infn

So let n become infinite and the relationship still holds.
 
Thank you, it's clear now.
 
Back
Top