Compactness Proof Strategies

Choosing a Compactness Proof Strategy

The Finite Intersection Property gives a useful way to turn compactness into an existence argument: if every finite collection of closed constraints can be satisfied inside a compact set, then all the constraints can be satisfied at once. That formulation is especially effective when the problem is about intersections. Other problems are more naturally phrased in terms of open covers or sequences. A good compactness proof begins by choosing the formulation that matches the conclusion.

Three familiar routes are available. The open-cover definition is suited to proving that some finite collection of neighborhoods suffices. The Finite Intersection Theorem is suited to proving that a family of closed constraints has a common point. The Sequential Characterization of Compactness in \(\mathbb{R}\) is suited to arguments that produce a sequence and then use a convergent subsequence. These are equivalent ways to use compactness in the real line, but they do not make every proof equally direct.

A common pattern is to find a quantity that varies continuously over a compact set. The Extreme Value Theorem then turns pointwise information into a uniform bound. We develop that pattern for distance to a closed set. It gives a precise separation result and, in turn, a useful method for working with open neighborhoods.

Distance to a Closed Set

For a nonempty set \(F\subseteq\mathbb{R}\), define the distance from \(x\in\mathbb{R}\) to \(F\) by taking the infimum of its distances to the points of \(F\). This definition does not require a point of \(F\) to be nearest to \(x\); in general, an infimum need not be attained.

The distance function changes by no more than the change in its input. This estimate is the reason it behaves well in compactness arguments.

Proof. Fix \(x,z\in\mathbb{R}\). For every \(y\in F\), the triangle inequality gives \(|x-y|\leq |x-z|+|z-y|\). Taking the infimum over \(y\in F\) gives \(d_F(x)\leq |x-z|+d_F(z)\). Interchanging \(x\) and \(z\) gives \(d_F(z)\leq |x-z|+d_F(x)\). Combining these inequalities yields \(|d_F(x)-d_F(z)|\leq |x-z|\). \(\square\)

In particular, \(d_F\) is continuous: given \(\varepsilon>0\), if \(|x-z|<\varepsilon\), then the lemma gives \(|d_F(x)-d_F(z)|<\varepsilon\). If \(F\) is also closed and \(x\notin F\), then \(d_F(x)>0\). Indeed, the complement of \(F\) is open, so some interval \((x-r,x+r)\), with \(r>0\), is disjoint from \(F\). Every \(y\in F\) then satisfies \(|x-y|\geq r\), and hence \(d_F(x)\geq r\).

Positive Separation from a Compact Set

Pointwise separation alone does not generally give a single positive lower bound that works at every point of a set. Compactness supplies the missing uniformity: the continuous distance function attains its minimum.

Proof. Since \(K\cap F=\varnothing\), every \(x\in K\) lies outside \(F\). The preceding observation shows that \(d_F(x)>0\) for every \(x\in K\). The distance function \(d_F\) is continuous by the Lipschitz lemma. Since \(K\) is nonempty and compact, the Extreme Value Theorem gives a point \(x_0\in K\) at which \(d_F\) attains its minimum on \(K\). Set \(m=d_F(x_0)\). Because \(d_F(x_0)>0\), we have \(m>0\), and \(d_F(x)\geq m\) for every \(x\in K\).

For any \(x\in K\) and \(y\in F\), the definition of infimum gives \(|x-y|\geq d_F(x)\geq m\). Thus \(\operatorname{dist}(K,F)\geq m>0\). Conversely, for \(x_0\in K\), the infimum of \(|x_0-y|\) over \(y\in F\) is \(d_F(x_0)=m\), so \(\operatorname{dist}(K,F)\leq m\). Therefore \(\operatorname{dist}(K,F)=m>0\), as required. \(\square\)

The roles of the hypotheses matter. Closedness of \(F\) ensures that each point of \(K\) has positive distance from \(F\); compactness of \(K\) turns those pointwise positive values into a positive minimum. Closedness of both sets, without compactness, does not guarantee positive distance.

Worked Examples: Putting the Strategy to Work

A Uniform Neighborhood Theorem

The same proof strategy gives a useful statement about open sets. If a compact set lies inside an open set, then it does not merely have a separate, possibly tiny neighborhood around each of its points. There is one positive radius that works throughout the compact set.

Proof. If \(U=\mathbb{R}\), any positive \(\varepsilon\) works. Otherwise, \(F=\mathbb{R}\setminus U\) is nonempty and closed. Since \(K\subseteq U\), the sets \(K\) and \(F\) are disjoint. By the Positive Separation Theorem, \(\delta=\operatorname{dist}(K,F)>0\). Set \(\varepsilon=\delta\). Suppose \(k\in K\), \(|x-k|<\varepsilon\), and \(x\notin U\). Then \(x\in F\), so the definition of \(\delta\) gives \(|x-k|\geq\delta=\varepsilon\), contradicting \(|x-k|<\varepsilon\). Therefore \(x\in U\), as claimed. \(\square\)

The strict inequality in the neighborhood condition is important: points at distance exactly \(\delta\) from \(K\) may lie in the complement of \(U\). Taking any smaller radius would also work, but the proof already gives the stated conclusion for distances strictly less than \(\delta\).

A Practical Checklist

When planning a compactness proof, first identify what the conclusion asks for. A common point suggests closed sets and the Finite Intersection Theorem. A finite collection of neighborhoods suggests the open-cover definition. A limiting point suggests a sequence and the Sequential Characterization of Compactness in \(\mathbb{R}\). In each case, verify the hypotheses before invoking compactness: in the separation theorem, for example, it is not enough to know that the sets are disjoint; the closedness and compactness assumptions each have a specific job.

Distance is one example of this general strategy. The key move was not to search separately for a lower bound at every point of \(K\), but to package those distances into a continuous function and then use compactness to obtain a minimum. When a proof requires a uniform estimate from pointwise information, that approach is often worth testing.