Write . If is a best approximation, then for every and real , the competitor satisfiesThe linear coefficient must be zero, since otherwise choosing sufficiently small of the appropriate sign makes this expression negative. Hence . In a real inner product space this is the desired orthogonality; in a complex inner product space, apply the same argument to to show that the imaginary part also vanishes. Thus for all .
Conversely, suppose this orthogonality holds. Every competitor has . The Pythagorean identity then givesEquality forces . We have proved the characterization, as well as uniqueness:This is the orthogonal projection onto a finite-dimensional subspace. Completeness of the ambient space is not needed. Existence follows, for example, by solving the invertible Gram matrix system in a basis of the finite-dimensional linear subspace.
The dual vectors are the Riesz dual basis in an inner product space, and lie in itself. Take the convention that the inner product is conjugate-linear in its first argument; in a real space the conjugations can simply be omitted. Since the form a basis, their Gram matrix is positive definite and invertible. Let and putThe proposed vectors satisfyThey therefore give the unique dual family within . Their own Gram matrix isThe inverse of a Hermitian positive-definite matrix is Hermitian, so . Using gives . ConsequentlyFor a real space this is the familiar identity . If the opposite complex inner-product convention is adopted, conjugate the coefficients in the dual-vector formula; the stated identity remains the same. Requiring the dual vectors to lie in is essential: arbitrary additions perpendicular to preserve the pairings with but alter the Gram matrix of the dual vectors.
Articles by others on the same topic
There are currently no matching articles.