POTW Preservation of Local Compactness

  • Thread starter Thread starter Euge
  • Start date Start date
  • Tags Tags
    Local
Click For Summary
The discussion centers on demonstrating that if a surjective continuous closed map of topological spaces has compact fibers, then the image space is locally compact if the domain is locally compact. Participants mention the relevance of saturated subsets in the context of quotient maps and their role in proofs related to fibrations and coverings. There is some confusion regarding the question's clarity, with requests for further explanation. The conversation highlights the importance of understanding topology definitions and their interconnections in forming a coherent argument. Overall, the focus remains on the implications of the properties of the map and the spaces involved.
Euge
Gold Member
MHB
POTW Director
Messages
2,072
Reaction score
245
Let ##f : X \to Y## be a surjective continuous closed map of topological spaces such that every fiber ##f^{-1}(y)## is compact. Show that ##Y## is locally compact if ##X## is locally compact.
 
Physics news on Phys.org
Doesnt this have to see with Saturated subsets of quotient maps, IIRC?
 
WWGD said:
Doesnt this have to see with Saturated subsets of quotient maps, IIRC?
I don't understand the question, could you please clarify?
 
Euge said:
I don't understand the question, could you please clarify?
Apologies, I remember the concept of saturated subsets playing a role in proofs when dealing with fibrations and coverings. Let me try to make the argument more precise.
 
I spent a long time on this before noticing you said f was a closed map! It doesn't seem like this has gotten a lot of traction, but it's really just chaining all the topology definitions together.

given ##y\in Y##, we need to find ##y\in U\subset K## with ##K## compact and ##U## open.

Since ##f## is subjective, there exists ##x\in X## such that ##f(x)=y##. ##X## is locally compact so there exists ##x\in V\subset L## with ##V## open and ##L## is compact.
##y\in f(V)\subset f(L)##. Since ##f## is continuous, ##f(L)## is compact. But ##f(V)## may not be open. The rest of the proof fixes this issue.

##X-V## is closed, and ##f## is a closed map, so ##f(X-V)## is closed. So ##Y-f(X-V)## is an open set, and it's a subset of ##f(L)##: if ##z\in Y-f(X-V)##, we know that ##f(a)=z## for some ##a##, and ##f(a)\notin f(X-V)## means ##a\in V## which we know is contained in ##L##. So we're close. But if ##f## is many to 1, this set may be too small, e.g. ##y## might be in ##f(X-V)##.

The final correction uses the compactness of $##f^{-1}(y)##. Let ##x_\alpha## be the pre image of ##y##, indexed by some set ##A## for notational convenience (the index set can literally be the pre image). For each ##\alpha##, ##x_\alpha \in V_\alpha \subset L_\alpha## with each ##V_\alpha## open and ##L_\alpha## compact. The ##V_\alpha##s are an open cover of ##f^{-1}(y)##, so there is a finite subcover, which is indexed by a finite set ##M\subset A##. ##X-\bigcup_M V_\alpha## is a closed set such that ##y\notin f(X-\bigcup_M V_\alpha)##. So ##y\in Y-f(X-\bigcup_M V_\alpha)## which is an open set, that is contained in ##f(\bigcup_M L_\alpha)##, by similar argument in the one index case. But ##\bigcup_M L_\alpha## is a finite union of compact sets so is compact, and then applying ##f## gives a compact subset of ##Y##.