logo

Compact Operator Equivalence Conditions 📂Banach Space

Compact Operator Equivalence Conditions

Theorem1

Let XX and YY be normed spaces. Let T:XYT : X \to Y be a linear operator. Then the following two propositions are equivalent.

  1. TT is a compact operator.
  2. TT maps “every bounded sequence in XX” to “a sequence in YY that has a convergent subsequence”.

Proof

1.2.1. \Longrightarrow 2.

Assume T:XYT : X \to Y is compact.

Let {xn}\left\{ x_{n} \right\} be a bounded sequence. By the definition of a compact operator, {Txn}\overline{\left\{ Tx_{n} \right\}} is compact, and because compactness in a metric space is equivalent to sequence compactness, {Txn}\left\{ Tx_{n} \right\} has a convergent subsequence.

Sequence Compactness

A metric space XX is sequentially compact if every sequence {xn}\left\{ x_{n} \right\} in XX has a subsequence {xnk}\left\{ x_{n_{k}} \right\} that converges to a point in XX.

2.1.2. \Longrightarrow 1.

Assume every bounded sequence {xn}\left\{ x_{n} \right\} has a subsequence {xnk}\left\{ x_{n_{k}} \right\} such that {Txnk}\left\{ Tx_{n_{k}} \right\} converges in YY.

Consider any bounded subset BXB \subset X and any sequence {yn}\left\{ y_{n} \right\} from T(B)T(B). Then, since BB is bounded, there exists a xnBx_{n} \in B such that {xn}\left\{ x_{n} \right\} is bounded. Therefore, by assumption, {Txn}\left\{ Tx_{n} \right\} has a convergent subsequence. Since {yn}={Txn}\left\{ y_{n} \right\} = \left\{ Tx_{n} \right\} was any sequence from T(B)T(B), every sequence from T(B)T(B) has a convergent subsequence. This means that T(B)T(B) is sequentially compact (= compact). But since initially BB was any bounded set, we obtain the result that TT maps any bounded set BB to a pre-compact set. Therefore, TT is a compact operator.

Corollary

The sum T1+T2T_{1} + T_{2} of two compact linear operators T1T_{1} and T2T_{2} is a compact operator. Moreover, for any constant α\alpha, αT1\alpha T_{1} is a compact operator.

Therefore, the set of compact linear operators from two normed spaces XX to YY forms a vector space.


  1. Erwin Kreyszig, Introductory Functional Analysis with Applications (1978), p ↩︎