TheoremBase

Proof

All sums below are finite sums in the field of real numbers, and ∣ ⋅ ∣|\,\cdot\,| is the absolute value. Write ∥ ⋅ ∥\lVert\,\cdot\,\rVert for the Euclidean norm. Unwinding Continuity at a Point for Maps Between Euclidean Spaces for a single coordinate function, fjf_j is continuous at aa precisely when for every real ε>0\varepsilon>0 there is a real δ>0\delta>0 such that every x∈Ex\in E with ∑i=1n(xi−ai)2<δ2\sum_{i=1}^{n}(x_i-a_i)^2<\delta^2 satisfies (fj(x)−fj(a))2<ε2(f_j(x)-f_j(a))^2<\varepsilon^2.

Necessity. Suppose ff is continuous at aa and fix jj. Let ε>0\varepsilon>0 be real and let δ>0\delta>0 be as in Continuity at a Point for Maps Between Euclidean Spaces for ff and ε\varepsilon. Let x∈Ex\in E satisfy ∑i=1n(xi−ai)2<δ2\sum_{i=1}^{n}(x_i-a_i)^2<\delta^2 and put z=f(x)−f(a)∈Rmz=f(x)-f(a)\in\mathbb{R}^m, so that zl=fl(x)−fl(a)z_l=f_l(x)-f_l(a) for every ll. Then ∑l=1mzl2<ε2\sum_{l=1}^{m}z_l^2<\varepsilon^2, and by claim 1 of Properties of the Euclidean Norm on Rn\mathbb{R}^n: Square, Distance, Vanishing, Coordinate Bounds, Homogeneity, Triangle Inequalities and Expansion of a Square this sum is ∥z∥2\lVert z\rVert^2, with 0≤∥z∥0\le\lVert z\rVert. Since 0<ε0<\varepsilon, claim 1 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field gives ∥z∥<ε\lVert z\rVert<\varepsilon, and claim 4 of Properties of the Euclidean Norm on Rn\mathbb{R}^n: Square, Distance, Vanishing, Coordinate Bounds, Homogeneity, Triangle Inequalities and Expansion of a Square gives ∣zj∣≤∥z∥|z_j|\le\lVert z\rVert, so ∣zj∣<ε|z_j|<\varepsilon and, by claim 1 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field again, ∣zj∣⋅∣zj∣<ε2|z_j|\cdot|z_j|<\varepsilon^2. Finally zj2≤∣zj2∣=∣zj∣⋅∣zj∣z_j^2\le|z_j^2|=|z_j|\cdot|z_j| by claims 3 and 4 of Properties of the Absolute Value in an Ordered Field, so (fj(x)−fj(a))2=zj2<ε2(f_j(x)-f_j(a))^2=z_j^2<\varepsilon^2. Hence fjf_j is continuous at aa.

Sufficiency. Suppose every fjf_j is continuous at aa. Put S=∑l=1m1S=\sum_{l=1}^{m}1. Every summand equals 11, which is nonnegative, so claim 6 of Properties of Finite Sums gives 1≤S1\le S; in particular 0<S0<S, and 0<S⋅S0<S\cdot S by claim 5 of Elementary Order Arithmetic in an Ordered Field.

Let ε>0\varepsilon>0 be real and put ε0=ε/(1+S)\varepsilon_0=\varepsilon/(1+S), a positive real number. For each l∈{1,…,m}l\in\{1,\dots,m\} choose, by the continuity of flf_l at aa, a real δl>0\delta_l>0 such that every x∈Ex\in E with ∑i=1n(xi−ai)2<δl2\sum_{i=1}^{n}(x_i-a_i)^2<\delta_l^2 satisfies (fl(x)−fl(a))2<ε02(f_l(x)-f_l(a))^2<\varepsilon_0^2. Let δ\delta be the least of δ1,…,δm\delta_1,\dots,\delta_m, a positive real number; since 0<δ≤δl0<\delta\le\delta_l, claim 2 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field gives δ2≤δl2\delta^2\le\delta_l^2 for every ll.

Let x∈Ex\in E satisfy ∑i=1n(xi−ai)2<δ2\sum_{i=1}^{n}(x_i-a_i)^2<\delta^2. Then that sum is smaller than δl2\delta_l^2 for every ll, so (fl(x)−fl(a))2<ε02(f_l(x)-f_l(a))^2<\varepsilon_0^2 for every ll. By claim 1 of Comparison and Absolute Value Bounds for Finite Sums of Real Numbers, applied with the summands (fl(x)−fl(a))2(f_l(x)-f_l(a))^2 and the constant summands ε02\varepsilon_0^2, and by the homogeneity of finite sums (claim 3 of Properties of Finite Sums),

∑l=1m(fl(x)−fl(a))2 ≤ ∑l=1mε02 = ε02 S.\sum_{l=1}^{m}\bigl(f_l(x)-f_l(a)\bigr)^2\ \le\ \sum_{l=1}^{m}\varepsilon_0^2\ =\ \varepsilon_0^2\,S .

Now (1+S)⋅(1+S)=1+S+S+S⋅S(1+S)\cdot(1+S)=1+S+S+S\cdot S, and 1+S+S⋅S1+S+S\cdot S is positive, so S<(1+S)⋅(1+S)S<(1+S)\cdot(1+S); multiplying by the positive number ε2/((1+S)(1+S))\varepsilon^2/\bigl((1+S)(1+S)\bigr) and using ε02=ε2/((1+S)(1+S))\varepsilon_0^2=\varepsilon^2/\bigl((1+S)(1+S)\bigr) gives ε02 S<ε2\varepsilon_0^2\,S<\varepsilon^2, the multiplication of a strict inequality by a positive number being claim 10 of Elementary Order Arithmetic in an Ordered Field. Hence

∑l=1m(fl(x)−fl(a))2<ε2,\sum_{l=1}^{m}\bigl(f_l(x)-f_l(a)\bigr)^2<\varepsilon^2,

which is the condition of Continuity at a Point for Maps Between Euclidean Spaces for ff at aa with the tolerance ε\varepsilon and the threshold δ\delta. Therefore ff is continuous at aa. ■\blacksquare

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…