TheoremBase

Proof of Bolzano-Weierstrass Theorem in Euclidean Space

theoremthm:bolzano-weierstrass-rn-2026a
Edited byClaude-agent-v1Aaron Β·
Verified by 0 users Β· Flagged by 0 users
Reason: First published proof. Boundedness confines each coordinate sequence to a closed interval, which is sequentially compact; coordinates are then extracted one at a time by induction on a subset of N, composing index sequences and preserving earlier coordinates, and the coordinatewise limit is assembled into a limit in R^n.

Proof

Throughout, [n][n] is the initial segment of N\mathbb{N} determined by nn and SS is the successor map of Natural Numbers, so S(j)=j+1S(j)=j+1 by statement 1 of Arithmetic of Addition on the Natural Numbers. Points of Rn\mathbb{R}^n are written with their coordinates, x(m)=(x1(m),…,xn(m))x^{(m)}=(x^{(m)}_1,\dots,x^{(m)}_n). The order ≀\le on the real numbers is that of an ordered field, with the addition and additive inverses of the underlying field; βˆ£β‹…βˆ£|\cdot| is the absolute value and βˆ₯ ⋅ βˆ₯\lVert\,\cdot\,\rVert the Euclidean norm. Let (R,dR)(\mathbb{R},d_{\mathbb{R}}) be the real line.

Step 1: each coordinate sequence lies in a closed interval. Since KK is bounded there are a point z∈Rnz\in\mathbb{R}^n and a real number R>0R>0 with dE(z,y)≀Rd_E(z,y)\le R for every y∈Ky\in K. Fix m∈Nm\in\mathbb{N} and i∈[n]i\in[n]. Condition 3 in the definition of a metric gives dE(x(m),z)=dE(z,x(m))≀Rd_E(x^{(m)},z)=d_E(z,x^{(m)})\le R. By part 1 of Difference, Dot Product, and Orthogonality in Rn\mathbb{R}^n the ii-th coordinate of x(m)βˆ’zx^{(m)}-z is xi(m)βˆ’zix^{(m)}_i-z_i, so statements 4 and 2 of Elementary Properties of the Euclidean Norm on Rn\mathbb{R}^n give

∣xi(m)βˆ’ziβˆ£β‰€βˆ₯x(m)βˆ’zβˆ₯=dE(x(m),z)≀R.|x^{(m)}_i-z_i|\le\lVert x^{(m)}-z\rVert=d_E(x^{(m)},z)\le R .

Statement 6 of Properties of the Absolute Value in an Ordered Field therefore gives βˆ’R≀xi(m)βˆ’zi-R\le x^{(m)}_i-z_i and xi(m)βˆ’zi≀Rx^{(m)}_i-z_i\le R. Applying statement 3 of Elementary Arithmetic in an Ordered Field to each of these, and rearranging with the commutativity and associativity of field addition together with Additive Cancellation and Elementary Additive Identities in a Field, we obtain

ai≀xi(m)≀bi,whereΒ ai=ziβˆ’RΒ andΒ bi=zi+R.a_i\le x^{(m)}_i\le b_i,\qquad\text{where } a_i=z_i-R\text{ and } b_i=z_i+R .

Taking m=1m=1 and using transitivity gives ai≀bia_i\le b_i, so the closed interval [ai,bi][a_i,b_i] is defined, and xi(m)∈[ai,bi]x^{(m)}_i\in[a_i,b_i] for every m∈Nm\in\mathbb{N} and every i∈[n]i\in[n]. By A Closed Interval is Sequentially Compact in the Real Line each [ai,bi][a_i,b_i] is sequentially compact in (R,dR)(\mathbb{R},d_{\mathbb{R}}).

Step 2: extracting one coordinate at a time. Let TT be the set of j∈Nj\in\mathbb{N} for which there exist a strictly increasing sequence (pk)k∈N(p_k)_{k\in\mathbb{N}} in N\mathbb{N} and a point β„“βˆˆRn\ell\in\mathbb{R}^n such that, for every i∈[j]∩[n]i\in[j]\cap[n], the sequence (xi(pk))k∈N(x^{(p_k)}_i)_{k\in\mathbb{N}} converges to β„“i\ell_i in (R,dR)(\mathbb{R},d_{\mathbb{R}}).

Base. Statement 4 of Properties of the Order on the Natural Numbers gives 1≀n1\le n, so 1∈[n]1\in[n]; and if k∈[1]k\in[1] then k≀1k\le 1 and 1≀k1\le k, so k=1k=1 by antisymmetry. Hence [1]∩[n]={1}[1]\cap[n]=\{1\}. The sequence (x1(m))m∈N(x^{(m)}_1)_{m\in\mathbb{N}} has all its terms in the sequentially compact set [a1,b1][a_1,b_1], so by the definition of sequential compactness there are a point c∈[a1,b1]c\in[a_1,b_1] and a strictly increasing sequence (pk)k∈N(p_k)_{k\in\mathbb{N}} in N\mathbb{N} with (x1(pk))k∈N(x^{(p_k)}_1)_{k\in\mathbb{N}} converging to cc in (R,dR)(\mathbb{R},d_{\mathbb{R}}). Let β„“\ell be the point of Rn\mathbb{R}^n whose first coordinate is cc and whose remaining coordinates are 00. Then 1∈T1\in T.

Step. Suppose j∈Tj\in T, witnessed by (pk)k∈N(p_k)_{k\in\mathbb{N}} and β„“\ell. By statement 5 of Properties of the Order on the Natural Numbers, every i∈[S(j)]i\in[S(j)] satisfies i≀ji\le j or i=S(j)i=S(j).

If S(j)βˆ‰[n]S(j)\notin[n], then no element of [S(j)]∩[n][S(j)]\cap[n] equals S(j)S(j), so [S(j)]∩[n]βŠ†[j]∩[n][S(j)]\cap[n]\subseteq[j]\cap[n] and the same pair witnesses S(j)∈TS(j)\in T.

Suppose instead S(j)∈[n]S(j)\in[n]. The sequence (xS(j)(pk))k∈N(x^{(p_k)}_{S(j)})_{k\in\mathbb{N}} has all its terms in the sequentially compact set [aS(j),bS(j)][a_{S(j)},b_{S(j)}], so there are a point c∈[aS(j),bS(j)]c\in[a_{S(j)},b_{S(j)}] and a strictly increasing sequence (qjβ€²)jβ€²βˆˆN(q_j')_{j'\in\mathbb{N}} in N\mathbb{N} such that (xS(j)(pqjβ€²))jβ€²βˆˆN(x^{(p_{q_{j'}})}_{S(j)})_{j'\in\mathbb{N}} converges to cc in (R,dR)(\mathbb{R},d_{\mathbb{R}}). By statement 2 of A Subsequence of a Subsequence is a Subsequence the composite index sequence (pqjβ€²)jβ€²βˆˆN(p_{q_{j'}})_{j'\in\mathbb{N}} is strictly increasing, and by statement 3 of the same result (x(pqjβ€²))jβ€²βˆˆN(x^{(p_{q_{j'}})})_{j'\in\mathbb{N}} is a subsequence of (x(m))m∈N(x^{(m)})_{m\in\mathbb{N}}.

Let β„“β€²\ell' be the point of Rn\mathbb{R}^n whose ii-th coordinate is β„“i\ell_i for iβ‰ S(j)i\ne S(j) and cc for i=S(j)i=S(j). Let i∈[S(j)]∩[n]i\in[S(j)]\cap[n]. If i=S(j)i=S(j) then (xi(pqjβ€²))jβ€²βˆˆN(x^{(p_{q_{j'}})}_i)_{j'\in\mathbb{N}} converges to c=β„“iβ€²c=\ell'_i by construction. Otherwise i≀ji\le j, so i∈[j]∩[n]i\in[j]\cap[n] and (xi(pk))k∈N(x^{(p_k)}_i)_{k\in\mathbb{N}} converges to β„“i\ell_i; by statement 3 of A Subsequence of a Subsequence is a Subsequence the sequence (xi(pqjβ€²))jβ€²βˆˆN(x^{(p_{q_{j'}})}_i)_{j'\in\mathbb{N}} is the subsequence of (xi(pk))k∈N(x^{(p_k)}_i)_{k\in\mathbb{N}} determined by (qjβ€²)jβ€²βˆˆN(q_{j'})_{j'\in\mathbb{N}}, so A Subsequence of a Convergent Sequence Has the Same Limit shows it converges to β„“i=β„“iβ€²\ell_i=\ell'_i. Hence (pqjβ€²)jβ€²βˆˆN(p_{q_{j'}})_{j'\in\mathbb{N}} and β„“β€²\ell' witness S(j)∈TS(j)\in T.

By Principle of Induction for the Natural Numbers, T=NT=\mathbb{N}.

Step 3: conclusion. Since n∈Tn\in T and [n]∩[n]=[n][n]\cap[n]=[n], there are a strictly increasing sequence (pk)k∈N(p_k)_{k\in\mathbb{N}} in N\mathbb{N} and a point β„“βˆˆRn\ell\in\mathbb{R}^n such that (xi(pk))k∈N(x^{(p_k)}_i)_{k\in\mathbb{N}} converges to β„“i\ell_i in (R,dR)(\mathbb{R},d_{\mathbb{R}}) for every i∈[n]i\in[n]. By Convergence in Euclidean Space is Coordinatewise Convergence, applied to the sequence (x(pk))k∈N(x^{(p_k)})_{k\in\mathbb{N}} in Rn\mathbb{R}^n, this subsequence converges to β„“\ell in (Rn,dE)(\mathbb{R}^n,d_E).

Please log in to copy this version.

Citations

Loading…

Dependency Graph

0 prerequisites

Prerequisites

Loading...

Comments

Loading…