Compactness — Mathematics Knowledge System

Overview

Compactness is a property of a topological space that makes it behave in many ways like a finite set, even when it has infinitely many points. The general (open-cover) definition says a space is compact if every collection of open sets that covers the space has a finite sub-collection that still covers it. For subsets of Euclidean space, this is equivalent to sequential compactness — every infinite sequence in the set has a subsequence converging to a point of the set — and, by the Heine–Borel theorem, to simply being closed and bounded. Just as every real-valued function on a finite set is bounded and attains its maximum and minimum, every continuous real-valued function on a compact space has these same properties (the extreme value theorem).

Intuition

Think of an open cover as covering a region with an arbitrary (possibly infinite) collection of overlapping 'patches' (open sets), each covering part of the space, together covering everything. Compactness says: no matter how the patches are chosen, you can always throw away all but finitely many of them and still have everything covered — the space can't 'require' infinitely many patches to stay covered. This captures the finite-set-like behavior: a finite set trivially has this property (there are only finitely many patches to begin with), and compactness generalizes it to infinite sets that are 'small' or 'complete' enough — like a closed bounded interval [a,b], where you truly cannot escape to infinity or sneak up on a missing boundary point.

Formal Definition

Definition

A topological space X is compact if every open cover of X admits a finite subcover:

\text{For every collection } \{U_\alpha\}_{\alpha \in A} \text{ of open sets with } X = \bigcup_{\alpha \in A} U_\alpha,\ \exists\, \alpha_1,\ldots,\alpha_n \in A \text{ such that } X = \bigcup_{i=1}^{n} U_{\alpha_i}

Every open cover — however large or infinite — has some finite sub-collection that still covers all of X

Open-cover definition of compactness

\text{(Heine–Borel) A subset } K \subseteq \mathbb{R}^n \text{ is compact} \iff K \text{ is closed and bounded.}

In Euclidean space specifically, the abstract open-cover property reduces to the concrete, checkable conditions of closedness and boundedness

Heine–Borel theorem

Notation

Notation	Meaning
$\{U_\alpha\}_{\alpha \in A}$	A collection of open sets whose union contains (covers) the space or set in question
$\{U_{\alpha_1}, \ldots, U_{\alpha_n}\}$	A finite sub-collection of a cover that still covers the whole space
$\text{every sequence has a convergent subsequence (limit in the set)}$	The sequence-based characterization of compactness, equivalent to the open-cover definition in metric spaces

Derivation

Sketch of why [0,1] is compact via the open-cover definition (the key step of the Heine–Borel theorem for a single interval), using a 'supremum of good points' argument.

\text{Let } \{U_\alpha\} \text{ be an open cover of } [0,1]. \text{ Let } S = \{x \in [0,1] : [0,x] \text{ has a finite subcover}\}.

Define S as the set of points up to which finite covering has succeeded

0 \in S \text{ (some } U_\alpha \text{ contains } 0\text{), so } S \neq \emptyset; \ S \text{ is bounded above by } 1.

S is nonempty and bounded, so by completeness of ℝ it has a supremum c = sup S

c \in [0,1] \implies c \in U_{\alpha_0} \text{ for some } \alpha_0 \text{, and } U_{\alpha_0} \text{ open} \implies (c-\delta, c+\delta) \subseteq U_{\alpha_0} \text{ for some } \delta>0.

The cover element containing c has room around c, since it's open

\text{Since } c=\sup S, \exists x \in S \cap (c-\delta, c]; [0,x] \text{ has a finite subcover, adding } U_{\alpha_0} \text{ covers } [0, \min(c+\delta,1)].

Extending the finite subcover of [0,x] by one more set (U_{α₀}) covers past c, forcing c=1 and showing 1∈S — completing the proof that [0,1] itself has a finite subcover

Properties

Continuous image of compact is compact

f: X \to Y \text{ continuous}, X \text{ compact} \implies f(X) \text{ compact}

Extreme Value Theorem

\text{If } f: K \to \mathbb{R} \text{ is continuous and } K \text{ is compact, then } f \text{ attains a maximum and a minimum value on } K.

Condition: A direct consequence of 'continuous image of compact is compact' applied to ℝ, where compact subsets are closed and bounded, hence contain their sup and inf.

Closed subsets of compact spaces are compact

K \text{ compact}, C \subseteq K \text{ closed} \implies C \text{ compact}

Compact subsets of Hausdorff spaces are closed

\text{If } X \text{ is Hausdorff and } K \subseteq X \text{ is compact, then } K \text{ is closed in } X.

Finite intersection property

X \text{ is compact} \iff \text{every family of closed sets with the finite intersection property has nonempty total intersection.}

Condition: An equivalent dual formulation of the open-cover definition, phrased via closed sets instead of open sets.

Applications

Compactness of configuration or phase spaces guarantees that physical quantities like energy, which are continuous functions on that space, attain actual maxima and minima rather than merely approaching a supremum that's never reached.

Worked Examples

Construct the open cover Uₙ = (1/n, 1) for n = 2, 3, 4, ..., whose union is all of (0,1).
$\bigcup_{n=2}^{\infty} \left(\frac{1}{n}, 1\right) = (0,1)$
Any finite sub-collection U_{n_1},...,U_{n_k} has union (1/N,1) where N = max(n₁,...,nₖ), which misses points near 0 (e.g. 1/(2N)).
$\bigcup_{i=1}^{k} U_{n_i} = \left(\frac{1}{N},1\right) \neq (0,1)$
Since no finite subcover exists for this particular cover, (0,1) fails the open-cover definition of compactness.
$\therefore (0,1) \text{ is not compact}$

Answer: (0,1) is not compact — it is bounded but not closed, consistent with Heine–Borel

Practice Problems

Difficulty 6/10

By the Heine–Borel theorem, which of the following subsets of ℝ is compact?

(0, 1)[0, ∞)[0, 1]ℝ

Difficulty 7/10

Prove that a compact subset K of a Hausdorff space X is closed.

Common Mistakes

Common Mistake

Assuming 'bounded' alone is enough to conclude a subset of ℝⁿ is compact.

Heine–Borel requires BOTH closed and bounded. (0,1) is bounded but not closed, and fails to be compact — as shown by the cover {(1/n,1)} having no finite subcover.

Common Mistake

Believing compactness is only about 'small' or literally finite sets.

Compact sets can be infinite (even uncountable) — [0,1] is compact and contains uncountably many points. Compactness is about every open cover admitting a FINITE subcover, not about the set itself being finite.

Quiz

By the Heine–Borel theorem, a subset of ℝⁿ is compact if and only if it is:

open and connectedclosed and boundedfinitean interval

Why does compactness matter for optimization (e.g. the Extreme Value Theorem)?

It makes functions linearA continuous function on a compact set always attains a maximum and minimumIt removes all constraintsIt guarantees a unique solution

Summary

A space is compact if every open cover has a finite subcover — the space cannot 'need' infinitely many open sets to stay covered.
In ℝⁿ, the Heine–Borel theorem gives a concrete equivalent: compact ⟺ closed and bounded.
Sequential compactness (every sequence has a convergent subsequence with limit in the set) is equivalent to compactness in metric spaces.
Continuous images of compact sets are compact, which is exactly why continuous functions on compact sets attain their maximum and minimum (Extreme Value Theorem).
Compact subsets of Hausdorff spaces are always closed, linking compactness back to the open-and-closed-sets framework.

References

WebsiteWikipedia — Compact space
BookMunkres, J. Topology, 2nd ed. Ch. 3.