TheoremBase

The extreme value theorem bounds the oscillation of ff, so the supremum of ∣f(x)−f(y)∣|f(x)-f(y)| over pairs at distance at most tt exists; this defines ω\omega. Monotonicity and domination are immediate from the definition, and the Heine-Cantor theorem gives the smallness of ω\omega near 00 required of a modulus of continuity.

Proof

Conventions. The order ≤\le and the arithmetic of R\mathbb{R} are those of the ordered field of real numbers. Two inequalities may be added: if a≤ba\le b and c≤dc\le d then a+c≤b+c≤b+da+c\le b+c\le b+d by the compatibility of ≤\le with addition and the transitivity of ≤\le.

Step 1: the oscillation of ff is bounded. Since ff is continuous on KK, for every x∈Kx\in K and every positive ε∈R\varepsilon\in\mathbb{R} there is a positive δ∈R\delta\in\mathbb{R} such that every y∈Ky\in K with d(x,y)<δd(x,y)<\delta satisfies dR(f(y),f(x))<εd_{\mathbb{R}}(f(y),f(x))<\varepsilon, that is, ∣f(y)−f(x)∣<ε|f(y)-f(x)|<\varepsilon. This is exactly the continuity property required in Extreme Value Theorem on a Compact Subset of a Metric Space, which therefore provides xmin⁡,xmax⁡∈Kx_{\min},x_{\max}\in K with

f(xmin⁡)≤f(z)≤f(xmax⁡)for every z∈K.f(x_{\min})\le f(z)\le f(x_{\max})\qquad\text{for every }z\in K .

Put C=f(xmax⁡)−f(xmin⁡)C=f(x_{\max})-f(x_{\min}). Let x,y∈Kx,y\in K. From f(xmin⁡)≤f(y)f(x_{\min})\le f(y) and claim 4 of Elementary Order Arithmetic in an Ordered Field we get −f(y)≤−f(xmin⁡)-f(y)\le-f(x_{\min}), and adding this to f(x)≤f(xmax⁡)f(x)\le f(x_{\max}) gives f(x)−f(y)≤Cf(x)-f(y)\le C. Exchanging xx and yy gives f(y)−f(x)≤Cf(y)-f(x)\le C, hence −C≤f(x)−f(y)-C\le f(x)-f(y) by claim 4 again, using −(f(y)−f(x))=f(x)−f(y)-\bigl(f(y)-f(x)\bigr)=f(x)-f(y). By claim 6 of Properties of the Absolute Value in an Ordered Field, ∣f(x)−f(y)∣≤C|f(x)-f(y)|\le C.

Step 2: construction of ω\omega. For t∈Tt\in T put

St={ ∣f(x)−f(y)∣ : x,y∈K with d(x,y)≤t }⊆R.S_{t}=\bigl\{\,|f(x)-f(y)|\ :\ x,y\in K\ \text{with}\ d(x,y)\le t\,\bigr\}\subseteq\mathbb{R}.

Since KK is nonempty, choose x0∈Kx_{0}\in K. Then d(x0,x0)=0≤td(x_{0},x_{0})=0\le t by the metric axioms, and ∣f(x0)−f(x0)∣=∣0∣=0|f(x_{0})-f(x_{0})|=|0|=0 by claim 1 of Properties of the Absolute Value in an Ordered Field, so 0∈St0\in S_{t} and StS_{t} is nonempty. By Step 1 the number CC is an upper bound for StS_{t}, so StS_{t} is bounded above, and Least Upper Bound Property of the Real Numbers provides a least upper bound of StS_{t}, which is unique by Uniqueness of the Supremum and of the Infimum. Define ω:T→R\omega:T\to\mathbb{R} by ω(t)=sup⁡St\omega(t)=\sup S_{t}.

Step 3: ω\omega is a modulus of continuity. Since 0∈St0\in S_{t} and ω(t)\omega(t) is an upper bound for StS_{t}, we have 0≤ω(t)0\le\omega(t) for every t∈Tt\in T, which is condition 1 of Modulus of Continuity.

For condition 2, let ε∈R\varepsilon\in\mathbb{R} be positive. By Heine-Cantor Theorem: Continuity on a Compact Metric Space or a Compact Subset Implies Uniform Continuity, applied with the metric spaces (M,d)(M,d) and (R,dR)(\mathbb{R},d_{\mathbb{R}}) and the compact set KK, the map ff is uniformly continuous on KK; hence there is a positive δ0∈R\delta_{0}\in\mathbb{R} such that all x,y∈Kx,y\in K with d(x,y)<δ0d(x,y)<\delta_{0} satisfy dR(f(x),f(y))=∣f(x)−f(y)∣<εd_{\mathbb{R}}(f(x),f(y))=|f(x)-f(y)|<\varepsilon. Put δ=δ0⋅2−1\delta=\delta_{0}\cdot2^{-1}, which by claim 8 of Elementary Order Arithmetic in an Ordered Field is positive and satisfies δ<δ0\delta<\delta_{0}. Let t∈Tt\in T satisfy t≤δt\le\delta, and let x,y∈Kx,y\in K satisfy d(x,y)≤td(x,y)\le t. Then d(x,y)≤δd(x,y)\le\delta and δ<δ0\delta<\delta_{0}, so d(x,y)<δ0d(x,y)<\delta_{0} by claim 2 of Elementary Order Arithmetic in an Ordered Field, whence ∣f(x)−f(y)∣<ε|f(x)-f(y)|<\varepsilon and in particular ∣f(x)−f(y)∣≤ε|f(x)-f(y)|\le\varepsilon. Thus ε\varepsilon is an upper bound for StS_{t}, and since ω(t)\omega(t) is the least upper bound of StS_{t} we get ω(t)≤ε\omega(t)\le\varepsilon. This is condition 2, so ω\omega is a modulus of continuity.

Step 4: claim 1. Let s,t∈Ts,t\in T with s≤ts\le t. If x,y∈Kx,y\in K satisfy d(x,y)≤sd(x,y)\le s, then d(x,y)≤td(x,y)\le t by transitivity; hence Ss⊆StS_{s}\subseteq S_{t}. Consequently ω(t)\omega(t), being an upper bound for StS_{t}, is an upper bound for SsS_{s}, and since ω(s)\omega(s) is the least upper bound of SsS_{s} we get ω(s)≤ω(t)\omega(s)\le\omega(t).

Step 5: claim 2. Let x,y∈Kx,y\in K and t∈Tt\in T satisfy d(x,y)≤td(x,y)\le t. Then ∣f(x)−f(y)∣∈St|f(x)-f(y)|\in S_{t}, and ω(t)\omega(t) is an upper bound for StS_{t}, so ∣f(x)−f(y)∣≤ω(t)|f(x)-f(y)|\le\omega(t).

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…