A linear operator between topological (often normed or Banach) vector spaces that maps bounded sets to relatively compact sets; equivalently, the image of any bounded sequence has a convergent subsequence in the target space.

Principle

Compactness encodes finite-dimensional approximation: compact operators can be approximated in operator norm by finite-rank operators in many common settings; they compress infinite-dimensional behavior into precompact images.

Demonstration

The integral operator K on L2 defined by (Kf)(x)=∫ k(x,y)f(y) dy with k∈L2 yields a compact operator under standard kernel regularity; the inclusion map from H1_0(Ω) into L2(Ω) is compact (Rellich embedding) while in finite dimensions every linear map is compact.

Misapplication

Assuming every bounded operator is compact (false in infinite dimensions); misusing spectral conclusions valid only for compact operators, e.g., claiming a bounded operator has a discrete point spectrum accumulating only at zero without verifying compactness.

Consequence

Compact operators enjoy specific spectral structure (point spectrum possibly accumulating only at zero), Fredholm alternative-type results, and better compactness-based convergence properties for sequences of images.

Reversal

A noncompact bounded operator fails to send bounded sets into relatively compact ones; typical examples include shift operators on ℓ2 or multiplication by an unbounded sequence, which exhibit continuous spectrum or lack finite-rank approximability in operator norm.

Boundary

Definition requires an ambient topology that makes relative compactness meaningful (normed or topological vector spaces); it excludes unbounded operators, maps whose images are bounded but not relatively compact, and properties that hold only pointwise or on particular subspaces.

Semantic Tension

‘Compact’ competes with ‘finite-rank’, ‘limit of finite-rank’, and with the broader notion of ‘completely continuous’ (an alias); the tension is resolved by specifying the topology and the mode of approximation (norm vs strong operator topology).

Synthesis

A compact operator is a linear transformation collapsing bounded sets to relatively compact ones so that infinite-dimensional behavior admits finite-dimensional approximations; this structural compression yields specific spectral and convergence consequences that distinguish compact operators from general bounded operators.