Proper map explained

In mathematics, a function between topological spaces is called proper if inverse images of compact subsets are compact. In algebraic geometry, the analogous concept is called a proper morphism.

Definition

There are several competing definitions of a "proper function". Some authors call a function

f:X\toY

between two topological spaces if the preimage of every compact set in

Y

is compact in

X.

Other authors call a map

f

if it is continuous and ; that is if it is a continuous closed map and the preimage of every point in

Y

is compact. The two definitions are equivalent if

Y

is locally compact and Hausdorff.Let

f:X\toY

be a closed map, such that

f-1(y)

is compact (in

X

) for all

y\inY.

Let

K

be a compact subset of

Y.

It remains to show that

f-1(K)

is compact.

Let

\left\{Ua:a\inA\right\}

be an open cover of

f-1(K).

Then for all

k\inK

this is also an open cover of

f-1(k).

Since the latter is assumed to be compact, it has a finite subcover. In other words, for every

k\inK,

there exists a finite subset

\gammak\subseteqA

such that

f-1(k)\subseteq

\cup
a\in\gammak

Ua.

The set

X\setminus

\cup
a\in\gammak

Ua

is closed in

X

and its image under

f

is closed in

Y

because

f

is a closed map. Hence the setV_k = Y \setminus f\left(X \setminus \cup_ U_\right) is open in

Y.

It follows that

Vk

contains the point

k.

Now

K\subseteq\cupkVk

and because

K

is assumed to be compact, there are finitely many points

k1,...,ks

such that

K\subseteq

s
\cup
i=1
V
ki

.

Furthermore, the set

\Gamma=

s
\cup
i=1
\gamma
ki
is a finite union of finite sets, which makes

\Gamma

a finite set.

Now it follows that

f-1(K)\subseteqf-1\left(

s
\cup
i=1
V
ki

\right)\subseteq\cupaUa

and we have found a finite subcover of

f-1(K),

which completes the proof.

If

X

is Hausdorff and

Y

is locally compact Hausdorff then proper is equivalent to . A map is universally closed if for any topological space

Z

the map

f x \operatorname{id}Z:X x Z\toY x Z

is closed. In the case that

Y

is Hausdorff, this is equivalent to requiring that for any map

Z\toY

the pullback

X x YZ\toZ

be closed, as follows from the fact that

X x YZ

is a closed subspace of

X x Z.

An equivalent, possibly more intuitive definition when

X

and

Y

are metric spaces is as follows: we say an infinite sequence of points

\{pi\}

in a topological space

X

if, for every compact set

S\subseteqX

only finitely many points

pi

are in

S.

Then a continuous map

f:X\toY

is proper if and only if for every sequence of points

\left\{pi\right\}

that escapes to infinity in

X,

the sequence

\left\{f\left(pi\right)\right\}

escapes to infinity in

Y.

Properties

f:X\toY

is called a if for every compact subset

K\subseteqY

there exists some compact subset

C\subseteqX

such that

f(C)=K.

f:X\toY

is a proper continuous map and

Y

is a compactly generated Hausdorff space (this includes Hausdorff spaces that are either first-countable or locally compact), then

f

is closed.[1]

Generalization

It is possible to generalize the notion of proper maps of topological spaces to locales and topoi, see .

References

Notes and References

  1. Palais. Richard S.. Richard Palais. When proper maps are closed. Proceedings of the American Mathematical Society. 1970. 24. 4. 835–836. 10.1090/s0002-9939-1970-0254818-x. free. 0254818 .