TheoremBase

Proof

Throughout, π\pi denotes the map described in claim 1, and coordinates of points of Rm+n\mathbb{R}^{m+n} are compared using the index decomposition recorded in the statement: every k∈[m+n]k\in[m+n] lies in [m][m] or equals m+jm+j for exactly one j∈[n]j\in[n], and not both.

Claim 1. Let w∈Rm+nw\in\mathbb{R}^{m+n} and let π(w)=(ξ,η)\pi(w)=(\xi,\eta) with ξk=wk\xi_k=w_k for k∈[m]k\in[m] and ηj=wm+j\eta_j=w_{m+j} for j∈[n]j\in[n]. Then ι(π(w))\iota(\pi(w)) and ww have the same coordinate at every index of [m+n][m+n], by the two defining clauses of ι\iota, so ι(π(w))=w\iota(\pi(w))=w. Conversely, for (ξ,η)∈Rm×Rn(\xi,\eta)\in\mathbb{R}^m\times\mathbb{R}^n the first entry of π(ι(ξ,η))\pi(\iota(\xi,\eta)) has kkth coordinate ι(ξ,η)k=ξk\iota(\xi,\eta)_k=\xi_k for k∈[m]k\in[m], and its second entry has jjth coordinate ι(ξ,η)m+j=ηj\iota(\xi,\eta)_{m+j}=\eta_j for j∈[n]j\in[n]; hence π(ι(ξ,η))=(ξ,η)\pi(\iota(\xi,\eta))=(\xi,\eta). So ι\iota and π\pi are mutually inverse, and ι\iota is a bijection.

Claim 2. Two points of Rm+n\mathbb{R}^{m+n} are equal exactly when they agree at every index of [m+n][m+n], so it suffices to compare coordinates at k∈[m]k\in[m] and at m+jm+j with j∈[n]j\in[n].

By Sum of Points of Rn\mathbb{R}^n the kkth coordinate of ι(ξ,η)+ι(ξ′,η′)\iota(\xi,\eta)+\iota(\xi',\eta') is ι(ξ,η)k+ι(ξ′,η′)k\iota(\xi,\eta)_k+\iota(\xi',\eta')_k. For k∈[m]k\in[m] this is ξk+ξk′\xi_k+\xi'_k, which is the kkth coordinate of ξ+ξ′\xi+\xi' and hence of ι(ξ+ξ′,η+η′)\iota(\xi+\xi',\eta+\eta'); for the index m+jm+j with j∈[n]j\in[n] it is ηj+ηj′\eta_j+\eta'_j, which is the jjth coordinate of η+η′\eta+\eta' and hence the (m+j)(m+j)th coordinate of ι(ξ+ξ′,η+η′)\iota(\xi+\xi',\eta+\eta'). This proves the first identity. The second is identical, using Scalar Multiple of a Point of Rn\mathbb{R}^n and that the kkth coordinate of λz\lambda z is λzk\lambda z_k; the third is identical, using the coordinatewise description of the difference in Difference, Dot Product, and Orthogonality in Rn\mathbb{R}^n.

Claim 3. Let c:[m+n]→Rc:[m+n]\to\mathbb{R} be the family ck=ι(ξ,η)k ι(ξ′,η′)kc_k=\iota(\xi,\eta)_k\,\iota(\xi',\eta')_k, so that by Difference, Dot Product, and Orthogonality in Rn\mathbb{R}^n

ι(ξ,η)⋅ι(ξ′,η′)=∑k=1m+nck.\iota(\xi,\eta)\cdot\iota(\xi',\eta')=\sum_{k=1}^{m+n}c_k .

By Splitting a Finite Sum at an Index this equals ∑k=1mck′+∑j=1ncj′′\sum_{k=1}^{m}c'_k+\sum_{j=1}^{n}c''_j, where c′c' is the restriction of cc to [m][m] and cj′′=cm+jc''_j=c_{m+j}. For k∈[m]k\in[m] we have ck′=ξkξk′c'_k=\xi_k\xi'_k, so ∑k=1mck′=ξ⋅ξ′\sum_{k=1}^{m}c'_k=\xi\cdot\xi'; and for j∈[n]j\in[n] we have cj′′=ηjηj′c''_j=\eta_j\eta'_j, so ∑j=1ncj′′=η⋅η′\sum_{j=1}^{n}c''_j=\eta\cdot\eta'. This proves the first identity of claim 3.

Taking ξ′=ξ\xi'=\xi and η′=η\eta'=\eta and using claim 1 of Elementary Properties of the Euclidean Norm on Rn\mathbb{R}^n, which gives ∥z∥2=z⋅z\lVert z\rVert^{2}=z\cdot z in each Euclidean space, yields ∥ι(ξ,η)∥2=∥ξ∥2+∥η∥2\lVert\iota(\xi,\eta)\rVert^{2}=\lVert\xi\rVert^{2}+\lVert\eta\rVert^{2}.

Claim 4. By claim 2 of Elementary Properties of the Euclidean Norm on Rn\mathbb{R}^n, dE(u,u′)=∥u−u′∥d_E(u,u')=\lVert u-u'\rVert in each Euclidean space. By claim 2 of the present lemma, ι(z)−ι(z′)=ι(ξ−ξ′,η−η′)\iota(z)-\iota(z')=\iota(\xi-\xi',\eta-\eta'), so by claim 3,

dE(ι(z),ι(z′))2=∥ι(ξ−ξ′,η−η′)∥2=∥ξ−ξ′∥2+∥η−η′∥2=dE(ξ,ξ′)2+dE(η,η′)2,d_E\bigl(\iota(z),\iota(z')\bigr)^{2}=\lVert\iota(\xi-\xi',\eta-\eta')\rVert^{2}=\lVert\xi-\xi'\rVert^{2}+\lVert\eta-\eta'\rVert^{2}=d_E(\xi,\xi')^{2}+d_E(\eta,\eta')^{2},

which is the first assertion.

Write α=dE(ξ,ξ′)\alpha=d_E(\xi,\xi'), β=dE(η,η′)\beta=d_E(\eta,\eta'), γ=dE(ι(z),ι(z′))\gamma=d_E(\iota(z),\iota(z')) and μ=d×(z,z′)\mu=d_{\times}(z,z'), all nonnegative because a metric takes nonnegative values. By Product Metric on the Cartesian Product of Two Metric Spaces, μ\mu is the maximum of α\alpha and β\beta, so α≤μ\alpha\le\mu and β≤μ\beta\le\mu by claim 1 of Elementary Properties of the Maximum of Two Elements, and μ\mu equals α\alpha or β\beta by claim 2 of that lemma.

Since μ\mu is one of α,β\alpha,\beta, we get μ2≤α2+β2=γ2\mu^{2}\le\alpha^{2}+\beta^{2}=\gamma^{2}, using that the other square is nonnegative by Nonnegativity of Squares in an Ordered Field and claim 3 of Elementary Arithmetic in an Ordered Field. As μ\mu and γ\gamma are nonnegative, claim 2 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field gives μ≤γ\mu\le\gamma.

For the upper bound, α≤μ\alpha\le\mu and β≤μ\beta\le\mu give α2≤μ2\alpha^{2}\le\mu^{2} and β2≤μ2\beta^{2}\le\mu^{2} by claim 2 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field, so

γ2=α2+β2≤μ2+μ2≤μ2+(μ μ+μ μ)+μ2=(μ+μ)2,\gamma^{2}=\alpha^{2}+\beta^{2}\le\mu^{2}+\mu^{2}\le \mu^{2}+\bigl(\mu\,\mu+\mu\,\mu\bigr)+\mu^{2}=(\mu+\mu)^{2},

the middle inequality holding because μ μ\mu\,\mu is nonnegative, and the last equality being field arithmetic. Since γ\gamma and μ+μ\mu+\mu are nonnegative, claim 2 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field gives γ≤μ+μ\gamma\le\mu+\mu.

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…