TheoremBase

The subset clause is read off the definitions; the continuity clauses follow by the epsilon/3 argument, comparing f with a single fKf_K chosen from uniform convergence and using the continuity or uniform continuity of fKf_K.

Proof

Preliminaries. For s,t,u∈Rs,t,u\in\mathbb{R}, ∣s−t∣≤∣s−u∣+∣u−t∣|s-t|\le|s-u|+|u-t| by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (xi); and ∣s−t∣=∣t−s∣|s-t|=|t-s|. Since dR(g(y),g(z))=∣g(y)−g(z)∣d_{\mathbb{R}}(g(y),g(z))=|g(y)-g(z)| for a map g:A→Rg:A\to\mathbb{R} and y,z∈Ay,z\in A (The Distance on the Real Numbers Is a Map with Values the Absolute Differences (i)), the condition of Continuous Map Between Metric Spaces (i) for gg to be continuous at x∈Ax\in A relative to AA reads: for every real ε>0\varepsilon>0 there is a real δ>0\delta>0 such that every y∈Ay\in A with dX(x,y)<δd_{X}(x,y)<\delta satisfies ∣g(y)−g(x)∣<ε|g(y)-g(x)|<\varepsilon; and that of Uniformly Continuous Map Between Metric Spaces (i) reads likewise with ∣g(y)−g(z)∣<ε|g(y)-g(z)|<\varepsilon for all y,z∈Ay,z\in A with dX(y,z)<δd_{X}(y,z)<\delta.

A third of ε\varepsilon. For a real ε>0\varepsilon>0 let η=(ε/2)/2\eta=(\varepsilon/2)/2. Two applications of Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (vii) give 0<η<ε/2<ε0<\eta<\varepsilon/2<\varepsilon, η+η=ε/2\eta+\eta=\varepsilon/2 and ε/2+ε/2=ε\varepsilon/2+\varepsilon/2=\varepsilon. If α,β,γ∈R\alpha,\beta,\gamma\in\mathbb{R} satisfy α<η\alpha<\eta, β<η\beta<\eta and γ<η\gamma<\eta, then α+β<η+η=ε/2\alpha+\beta<\eta+\eta=\varepsilon/2 by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (ii), γ<ε/2\gamma<\varepsilon/2 by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (i), and so α+β+γ<ε/2+ε/2=ε\alpha+\beta+\gamma<\varepsilon/2+\varepsilon/2=\varepsilon by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (ii).

Proof of (i). Let T⊆S⊆AT\subseteq S\subseteq A, and suppose that (fk)k∈N(f_{k})_{k\in\mathbb{N}} converges uniformly to ff on SS. Let ε>0\varepsilon>0 be real. By Pointwise and Uniform Convergence of a Sequence of Real-Valued Maps (ii) there is K∈NK\in\mathbb{N} with ∣fk(x)−f(x)∣<ε|f_{k}(x)-f(x)|<\varepsilon for every k∈Nk\in\mathbb{N} with k≥Kk\ge K and every x∈Sx\in S. Every x∈Tx\in T lies in SS, so the same KK serves for TT; as ε\varepsilon was arbitrary, (fk)k∈N(f_{k})_{k\in\mathbb{N}} converges uniformly to ff on TT by Pointwise and Uniform Convergence of a Sequence of Real-Valued Maps (ii). For a fixed x∈Sx\in S the same KK gives ∣fk(x)−f(x)∣<ε|f_{k}(x)-f(x)|<\varepsilon for every k≥Kk\ge K; as ε\varepsilon was arbitrary, (fk(x))k∈N(f_{k}(x))_{k\in\mathbb{N}} converges to f(x)f(x) by Convergent Sequences of Real Numbers (i), and as x∈Sx\in S was arbitrary, (fk)k∈N(f_{k})_{k\in\mathbb{N}} converges pointwise to ff on SS by Pointwise and Uniform Convergence of a Sequence of Real-Valued Maps (i).

Proof of (ii). Let xx, rr and SS be as there. First, x∈Sx\in S: indeed dX(x,x)=0<rd_{X}(x,x)=0<r by Metric Space (iii), so x∈BdX(x,r)x\in B_{d_{X}}(x,r) by Open Ball in a Metric Space (ii), and x∈Ax\in A, so x∈Sx\in S by the hypothesis on SS.

The choices are made in this order: a real ε>0\varepsilon>0 is given; then η=(ε/2)/2\eta=(\varepsilon/2)/2; then KK; then δ0\delta_{0}; then δ\delta. By Pointwise and Uniform Convergence of a Sequence of Real-Valued Maps (ii) applied to η\eta, there is K∈NK\in\mathbb{N} with ∣fk(y)−f(y)∣<η|f_{k}(y)-f(y)|<\eta for every k∈Nk\in\mathbb{N} with k≥Kk\ge K and every y∈Sy\in S; we use it for k=Kk=K only. Since fKf_{K} is continuous at xx relative to AA, there is a real δ0>0\delta_{0}>0 such that every y∈Ay\in A with dX(x,y)<δ0d_{X}(x,y)<\delta_{0} satisfies ∣fK(y)−fK(x)∣<η|f_{K}(y)-f_{K}(x)|<\eta. By Closure Facts in an Ordered Field: Negatives, Differences, Quotients, Absolute Values, Powers, Natural Numbers, Halves and Minima (viii) there is a real δ\delta with 0<δ0<\delta, δ≤δ0\delta\le\delta_{0} and δ≤r\delta\le r.

Let y∈Ay\in A with dX(x,y)<δd_{X}(x,y)<\delta. Then dX(x,y)<rd_{X}(x,y)<r and dX(x,y)<δ0d_{X}(x,y)<\delta_{0} by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (i). Hence y∈BdX(x,r)y\in B_{d_{X}}(x,r) by Open Ball in a Metric Space (ii), and as y∈Ay\in A, y∈Sy\in S; so ∣f(y)−fK(y)∣=∣fK(y)−f(y)∣<η|f(y)-f_{K}(y)|=|f_{K}(y)-f(y)|<\eta. As x∈Sx\in S, also ∣fK(x)−f(x)∣<η|f_{K}(x)-f(x)|<\eta, and ∣fK(y)−fK(x)∣<η|f_{K}(y)-f_{K}(x)|<\eta by the choice of δ0\delta_{0}. Two applications of the triangle inequality of the preliminaries give

∣f(y)−f(x)∣≤∣f(y)−fK(y)∣+∣fK(y)−f(x)∣≤∣f(y)−fK(y)∣+∣fK(y)−fK(x)∣+∣fK(x)−f(x)∣,|f(y)-f(x)|\le|f(y)-f_{K}(y)|+|f_{K}(y)-f(x)|\le|f(y)-f_{K}(y)|+|f_{K}(y)-f_{K}(x)|+|f_{K}(x)-f(x)|,

and the right-hand side is smaller than ε\varepsilon by the third of ε\varepsilon. As ε>0\varepsilon>0 was arbitrary, ff is continuous at xx relative to AA.

Proof of (iii). Let x∈Ax\in A. The real number 11 is positive by Rules of Arithmetic and Order in an Ordered Field (vii), and AA contains every point of BdX(x,1)B_{d_{X}}(x,1) that lies in AA. Each fkf_{k}, being continuous on AA, is continuous at xx relative to AA by Continuous Map Between Metric Spaces (ii), and (fk)k∈N(f_{k})_{k\in\mathbb{N}} converges uniformly to ff on AA. So (ii), applied with r=1r=1 and S=AS=A, shows that ff is continuous at xx relative to AA. As x∈Ax\in A was arbitrary, ff is continuous on AA by Continuous Map Between Metric Spaces (ii).

Proof of (iv). The choices are made in this order: a real ε>0\varepsilon>0 is given; then η=(ε/2)/2\eta=(\varepsilon/2)/2; then KK; then δ\delta. By Pointwise and Uniform Convergence of a Sequence of Real-Valued Maps (ii) applied to η\eta, there is K∈NK\in\mathbb{N} with ∣fk(y)−f(y)∣<η|f_{k}(y)-f(y)|<\eta for every k∈Nk\in\mathbb{N} with k≥Kk\ge K and every y∈Ay\in A; we use it for k=Kk=K only. Since fKf_{K} is uniformly continuous on AA, there is a real δ>0\delta>0 such that all u,v∈Au,v\in A with dX(u,v)<δd_{X}(u,v)<\delta satisfy ∣fK(u)−fK(v)∣<η|f_{K}(u)-f_{K}(v)|<\eta.

Let u,v∈Au,v\in A with dX(u,v)<δd_{X}(u,v)<\delta. Then ∣f(u)−fK(u)∣=∣fK(u)−f(u)∣<η|f(u)-f_{K}(u)|=|f_{K}(u)-f(u)|<\eta, ∣fK(u)−fK(v)∣<η|f_{K}(u)-f_{K}(v)|<\eta and ∣fK(v)−f(v)∣<η|f_{K}(v)-f(v)|<\eta, and two applications of the triangle inequality of the preliminaries give

∣f(u)−f(v)∣≤∣f(u)−fK(u)∣+∣fK(u)−fK(v)∣+∣fK(v)−f(v)∣<ε,|f(u)-f(v)|\le|f(u)-f_{K}(u)|+|f_{K}(u)-f_{K}(v)|+|f_{K}(v)-f(v)|<\varepsilon ,

the last step by the third of ε\varepsilon. As ε>0\varepsilon>0 was arbitrary, ff is uniformly continuous on AA.

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…