220 Complete Set

A set of vectors whose linear combinations can approximate every vector in a larger set that is used as a basis-like tool to represent other vectors.

definition (Complete Set) A subset of a normed space whose span is dense in that space, meaning any element in the space can be represented by a linear combination of the set’s members. The set must satisfy the following condition:

  • The only vector in the space that is orthogonal to every element in the set is the zero vector.

where

  • \(M\) is the set of vectors being tested for completeness.
  • \(H\) is the Hilbert space containing the set.
  • \(\text{span } M\) is the set of all finite linear combinations of elements in \(M\).
  • \(\delta(x - x')\) is the Dirac delta function.
  • \(1 = \sum |e_n\rangle\langle e_n|\) is the completeness relation for a discrete orthonormal basis.

Note:

  • the same notion applies in an inner product space.
  • \(M\) may be a set of functions.
  • an element may be approximated rather than exactly represented by a linear combination.
  • \(\delta(x - x')\) is used to express the completeness relation for sets with continuous indices.
  • the completeness relation is also called the resolution of the identity.

220.1 Elementary Example

220.1.1 Simple

A complete set spans a dense subspace. In \(\mathbb{R}^{2}\), the two standard basis vectors form a complete set.

\[ M = \{ e_{1},\ e_{2} \} \]

\[ e_{1} = (1,0),\quad e_{2} = (0,1) \]

\[ \operatorname{span} M = \mathbb{R}^{2} \]

where

  • \(M\) is the set being tested for completeness.
  • \(\operatorname{span} M\) is the set of finite linear combinations of members of \(M\).

220.1.2 General

In \(\mathbb{R}^{3}\), the three standard basis vectors form a complete set: only the zero vector is orthogonal to all three.

\[ M = \{ e_{1},\ e_{2},\ e_{3} \} \]

\[ \operatorname{span} M = \mathbb{R}^{3} \]

where

  • if \(\langle v, e_{i} \rangle = 0\) for \(i = 1,2,3\), then \(v = 0\).