ProofComplete

Hilbert Spaces - Key Proof

We present a detailed proof of the Projection Theorem, one of the most fundamental results in Hilbert space theory, which establishes the existence and uniqueness of best approximations in closed subspaces.

TheoremProjection Theorem

Let HH be a Hilbert space and M⊂HM \subset H a closed subspace. Then for every x∈Hx \in H, there exists a unique m∈Mm \in M such that ∥x−m∥=inf⁡y∈M∥x−y∥=d(x,M)\|x - m\| = \inf_{y \in M} \|x - y\| = d(x, M) Moreover, mm is characterized by the condition (x−m)⊥M(x - m) \perp M.

Proof

Existence: Let d=inf⁡y∈M∥x−y∥d = \inf_{y \in M} \|x - y\|. Choose a sequence (mn)(m_n) in MM such that ∥x−mn∥→d\|x - m_n\| \to d as n→∞n \to \infty.

We show (mn)(m_n) is Cauchy using the parallelogram law: ∥mn−mk∥2=∥(x−mk)−(x−mn)∥2\|m_n - m_k\|^2 = \|(x - m_k) - (x - m_n)\|^2 =2∥x−mk∥2+2∥x−mn∥2−∥2x−mk−mn∥2= 2\|x - m_k\|^2 + 2\|x - m_n\|^2 - \|2x - m_k - m_n\|^2 =2∥x−mk∥2+2∥x−mn∥2−4∥x−mk+mn2∥2= 2\|x - m_k\|^2 + 2\|x - m_n\|^2 - 4\left\|x - \frac{m_k + m_n}{2}\right\|^2

Since MM is a subspace, (mk+mn)/2∈M(m_k + m_n)/2 \in M, so ∥x−mk+mn2∥2≥d2\left\|x - \frac{m_k + m_n}{2}\right\|^2 \geq d^2

As n,k→∞n, k \to \infty, we have ∥x−mn∥2,∥x−mk∥2→d2\|x - m_n\|^2, \|x - m_k\|^2 \to d^2, therefore ∥mn−mk∥2≤2d2+2d2−4d2=0\|m_n - m_k\|^2 \leq 2d^2 + 2d^2 - 4d^2 = 0

So (mn)(m_n) is Cauchy. Since HH is complete and MM is closed, mn→m∈Mm_n \to m \in M. By continuity of the norm, ∥x−m∥=d\|x - m\| = d.

Orthogonality Characterization: For any y∈My \in M and λ∈K\lambda \in \mathbb{K}, we have m+λy∈Mm + \lambda y \in M, so ∥x−m∥2≤∥x−m−λy∥2=∥x−m∥2−2Re(λ⟨x−m,y⟩)+∣λ∣2∥y∥2\|x - m\|^2 \leq \|x - m - \lambda y\|^2 = \|x - m\|^2 - 2\text{Re}(\lambda \langle x - m, y \rangle) + |\lambda|^2 \|y\|^2

This gives 2Re(λ⟨x−m,y⟩)≤∣λ∣2∥y∥22\text{Re}(\lambda \langle x - m, y \rangle) \leq |\lambda|^2 \|y\|^2 for all λ\lambda.

Choosing λ=t⟨x−m,y⟩\lambda = t \langle x - m, y \rangle for small t>0t > 0 yields 2t∣⟨x−m,y⟩∣2≤t2∣⟨x−m,y⟩∣2∥y∥22t |\langle x - m, y \rangle|^2 \leq t^2 |\langle x - m, y \rangle|^2 \|y\|^2. Dividing by tt and letting t→0t \to 0 gives ⟨x−m,y⟩=0\langle x - m, y \rangle = 0 for all y∈My \in M.

Uniqueness: If m1,m2∈Mm_1, m_2 \in M both satisfy ∥x−m1∥=∥x−m2∥=d\|x - m_1\| = \|x - m_2\| = d, then by the parallelogram law: ∥m1−m2∥2=2∥x−m1∥2+2∥x−m2∥2−4∥x−m1+m22∥2\|m_1 - m_2\|^2 = 2\|x - m_1\|^2 + 2\|x - m_2\|^2 - 4\left\|x - \frac{m_1 + m_2}{2}\right\|^2 ≤2d2+2d2−4d2=0\leq 2d^2 + 2d^2 - 4d^2 = 0

Therefore m1=m2m_1 = m_2.

■
ExampleComputing Projections
  1. Finite-Dimensional Subspaces: If M=span{e1,…,en}M = \text{span}\{e_1, \ldots, e_n\} is orthonormal, then PM(x)=∑i=1n⟨x,ei⟩eiP_M(x) = \sum_{i=1}^n \langle x, e_i \rangle e_i

  2. Fourier Series: The projection of f∈L2[0,2π]f \in L^2[0, 2\pi] onto trigonometric polynomials of degree ≤N\leq N is the partial Fourier sum

Remark

The Projection Theorem requires the subspace MM to be closed. For non-closed subspaces, the infimum may not be attained. This is why completeness of Hilbert spaces is essential.

This theorem is the foundation for least squares approximation, Fourier analysis, and variational methods in PDEs.