The key to many computations to follow in this chapter can be found in Exercise I.45.
Prove that in an inner product space, .
For any and in a normed linear space , .
For any and in a normed linear space , if then .
For an element and a subset of a normed linear space , the distance from to in , written , is .
For a closed convex (non-empty) subset of a Hilbert space and a vector in , there exists a vector in such that for all in distinct from . That is, there exists a unique point in nearest to .
Proof. Let be a sequence of elements of such that converges to . We claim that is Cauchy. Indeed, using Theorem I.46, Note that is in because is convex. So and it follows that is Cauchy.
As a Hilbert space, is complete, so the sequence converges to some element of . Because is closed, . Thus , and for any element of other than , So and .
Let be a closed subspace of a Hilbert space . The orthogonal projection of an element on is the point in nearest to . The operator that maps each element of to its orthogonal projection on is called the projection of on .
Show that the projection of a Hilbert space onto a closed subspace is a linear operator.
For an inner product space , elements and are orthogonal, denoted , if . .
Let be a subset of an inner product space . The orthogonal complement of in , written , is the set of all vectors in such that for all vectors in .
Prove that for any subset of an inner product space , is a subspace of .
Prove that for any subset of an inner product space , .
Prove that for any subspace of an inner product space , .
For any subset of an inner product space , is a closed subspace of .
Proof. The subset is a subspace by Exercise I.54. To show that it is closed, consider a sequence of elements of that converges in to a vector . Then using Proposition I.3, for all in . Taking limits on each side completes the proof.
A vector space is the direct sum of two subspaces and , which we write as , if each vector in can be written uniquely as the sum of a vector in and a vector in .
For any closed subspace of a Hilbert space , we have .
Proof. Take an element in the Hilbert space . By Theorem I.49 there exists a point in nearest to . In particular, . We claim the orthogonal decomposition we seek is .
We must first show that is in . Indeed, for and , since is in . So, for , . Dividing by gives . Thus for all . Repeating the same argument with in place of leads to the conclusion for all , meaning, in fact, for all . Repeating the above with and shows that for all . Thus, is in .
By Exercise I.56, the decomposition is unique (since two different decompositions would give you something non-zero in ). Thus .
For a subset of an inner product space , .
For a subset of a Hilbert space , .
Proof. Let so that is a closed subspace of . We claim that . Indeed, by Exercise I.55. Using Theorem I.59, for , write where and . Then , and so So and thus . Hence , so , and thus . Finally, since by Exercise I.60, we have .
Previous: Completeness | back to top | Next: Finite Dimensional Spaces |