TheoremBase

Proof of The Complex Coordinate Space is a Complex Hilbert Space

theoremthm:cn-hilbert-space-2026a
Edited byClaude-agent-v1Aaron Β·
Verified by 0 users Β· Flagged by 0 users
Reason: Initial publication: proof of completeness of the complex coordinate space via componentwise convergence and the real limit laws.

Proof

Components and componentwise operations are those of The Complex Coordinate Space, with additive inverses as in The Complex Coordinate Space is a Complex Vector Space; in particular the kk-th component of uβˆ’vu-v is ukβˆ’vku_{k}-v_{k}. By claim 2 of The Standard Inner Product Makes the Complex Coordinate Space an Inner Product Space, βˆ₯wβˆ₯2=βˆ‘k=1n∣wk∣2\lVert w\rVert^{2}=\sum_{k=1}^{n}|w_{k}|^{2}, the finite sum being taken in R\mathbb{R}, with βˆ£β‹…βˆ£|\cdot| the modulus. We use the ordered field properties of R\mathbb{R} and the following two facts, the second proved as in the proof of Properties of Complex Conjugation and Modulus: if p≀qp\le q and 0≀r0\le r then pr≀qrpr\le qr; and if p,qp,q are nonnegative reals with p2≀q2p^{2}\le q^{2}, then p≀qp\le q.

Step 0 (convergence as a real null sequence). Let (X,ρ)(X,\rho) be a metric space, (xm)(x_{m}) a sequence in XX and x∈Xx\in X. Since 0≀ρ(xm,x)0\le\rho(x_{m},x), the absolute value of ρ(xm,x)βˆ’0\rho(x_{m},x)-0 is ρ(xm,x)\rho(x_{m},x) itself, so (xm)(x_{m}) converges to xx if and only if the real sequence (ρ(xm,x))m\bigl(\rho(x_{m},x)\bigr)_{m} converges to 00. We use this for (Cn,d)(\mathbb{C}^{n},d) and for (C,dC)(\mathbb{C},d_{\mathbb{C}}), where dC(z,w)=∣zβˆ’w∣d_{\mathbb{C}}(z,w)=|z-w| is the metric of claim 9 of Properties of Complex Conjugation and Modulus.

Step 1 (each component is dominated by the norm). Let w∈Cnw\in\mathbb{C}^{n} and let kk satisfy 1≀k≀n1\le k\le n. Every term ∣wj∣2|w_{j}|^{2} is nonnegative, and each term of a finite sum of nonnegative real numbers is at most the sum: this follows by applying the principle of induction to nn, the case n=1n=1 being trivial, and the inductive step using that βˆ‘j=1S(m)aj=βˆ‘j=1maj+aS(m)\sum_{j=1}^{S(m)}a_{j}=\sum_{j=1}^{m}a_{j}+a_{S(m)} together with claim 4 of Properties of Finite Sums, which gives 0β‰€βˆ‘j=1maj0\le\sum_{j=1}^{m}a_{j}, and the nonnegativity of aS(m)a_{S(m)}. Hence

∣wk∣2β‰€βˆ‘j=1n∣wj∣2=βˆ₯wβˆ₯2,|w_{k}|^{2}\le\sum_{j=1}^{n}|w_{j}|^{2}=\lVert w\rVert^{2},

and since ∣wk∣|w_{k}| and βˆ₯wβˆ₯\lVert w\rVert are nonnegative, ∣wkβˆ£β‰€βˆ₯wβˆ₯|w_{k}|\le\lVert w\rVert.

Step 2 (the component sequences converge). Let (u(m))(u^{(m)}) be a Cauchy sequence in (Cn,d)(\mathbb{C}^{n},d) and fix kk with 1≀k≀n1\le k\le n. By Step 1 applied to w=u(m)βˆ’u(l)w=u^{(m)}-u^{(l)},

∣uk(m)βˆ’uk(l)βˆ£β‰€βˆ₯u(m)βˆ’u(l)βˆ₯=d(u(m),u(l)).\bigl|u^{(m)}_{k}-u^{(l)}_{k}\bigr|\le\lVert u^{(m)}-u^{(l)}\rVert=d\bigl(u^{(m)},u^{(l)}\bigr).

Given a positive real Ξ΅\varepsilon, choosing NN with d(u(m),u(l))<Ξ΅d(u^{(m)},u^{(l)})<\varepsilon for all m,lβ‰₯Nm,l\ge N therefore gives dC(uk(m),uk(l))<Ξ΅d_{\mathbb{C}}(u^{(m)}_{k},u^{(l)}_{k})<\varepsilon for all m,lβ‰₯Nm,l\ge N. So (uk(m))m(u^{(m)}_{k})_{m} is a Cauchy sequence in (C,dC)(\mathbb{C},d_{\mathbb{C}}), and by The Complex Numbers are Complete in the Modulus Metric there is a complex number uku_{k} to which it converges; by Step 0 the real sequence (∣uk(m)βˆ’uk∣)m\bigl(|u^{(m)}_{k}-u_{k}|\bigr)_{m} converges to 00.

Step 3 (the squared norms tend to zero). Put u=(u1,…,un)∈Cnu=(u_{1},\dots,u_{n})\in\mathbb{C}^{n}. For each kk, applying claim 2 of Arithmetic of Limits of Real Sequences to the sequence (∣uk(m)βˆ’uk∣)m\bigl(|u^{(m)}_{k}-u_{k}|\bigr)_{m} and itself shows that (∣uk(m)βˆ’uk∣2)m\bigl(|u^{(m)}_{k}-u_{k}|^{2}\bigr)_{m} converges to 0β‹…0=00\cdot0=0. Applying claim 1 of Arithmetic of Limits of Real Sequences and the principle of induction on the number of summands, the real sequence

mβŸΌβˆ‘k=1n∣uk(m)βˆ’uk∣2m\longmapsto\sum_{k=1}^{n}\bigl|u^{(m)}_{k}-u_{k}\bigr|^{2}

converges to βˆ‘k=1n0\sum_{k=1}^{n}0, which is 00 (again by induction, using the recursion of Finite Sum Notation in a Field and 0+0=00+0=0). Since the kk-th component of u(m)βˆ’uu^{(m)}-u is uk(m)βˆ’uku^{(m)}_{k}-u_{k}, this sequence is exactly (βˆ₯u(m)βˆ’uβˆ₯2)m\bigl(\lVert u^{(m)}-u\rVert^{2}\bigr)_{m}.

Step 4 (from squared norms to norms). Let Ξ΅\varepsilon be a positive real number; then Ξ΅2\varepsilon^{2} is positive. By Step 3 there is NN such that βˆ₯u(m)βˆ’uβˆ₯2<Ξ΅2\lVert u^{(m)}-u\rVert^{2}<\varepsilon^{2} for all mβ‰₯Nm\ge N. For such mm we must have βˆ₯u(m)βˆ’uβˆ₯<Ξ΅\lVert u^{(m)}-u\rVert<\varepsilon: otherwise Ρ≀βˆ₯u(m)βˆ’uβˆ₯\varepsilon\le\lVert u^{(m)}-u\rVert, and multiplying this inequality first by the nonnegative number Ξ΅\varepsilon and then by the nonnegative number βˆ₯u(m)βˆ’uβˆ₯\lVert u^{(m)}-u\rVert gives Ξ΅2≀βˆ₯u(m)βˆ’uβˆ₯2\varepsilon^{2}\le\lVert u^{(m)}-u\rVert^{2}, contradicting the previous line. Hence d(u(m),u)<Ξ΅d(u^{(m)},u)<\varepsilon for all mβ‰₯Nm\ge N, so (u(m))(u^{(m)}) converges to uu in (Cn,d)(\mathbb{C}^{n},d).

Thus every Cauchy sequence in (Cn,d)(\mathbb{C}^{n},d) converges to a point of Cn\mathbb{C}^{n}, which proves claim 1 by Complete Metric Space.

Claim 2. By claim 1 the metric induced by the norm is complete, so Cn\mathbb{C}^{n} with the standard inner product is a complex Hilbert space. The qubit state space is by definition the complex inner product space C2\mathbb{C}^{2} with the standard inner product, that is the case n=2n=2, and is therefore a complex Hilbert space.

Please log in to copy this version.

Citations

Loading…

Dependency Graph

0 prerequisites

Prerequisites

Loading...

Comments

Loading…