TheoremBase

Direct epsilon-delta arguments: delta equal to epsilon for the identity, the two moduli chained for a composition, and for the reciprocal a lower bound |g(t)| > |g(x)|/2 near x combined with the identity 1/g(t) - 1/g(x) = (g(x) - g(t)) / (g(t) g(x)).

Proof

Proof of (i). Let ι:A→X\iota:A\to X be the map u↦uu\mapsto u, so that ι(u)=u\iota(u)=u for u∈Au\in A (Sets and Maps: Ordinary Notation (v)). Let x∈Ax\in A and let ε>0\varepsilon>0 be real; put Ī“=ε\delta=\varepsilon. Every y∈Ay\in A with d(x,y)<Ī“d(x,y)<\delta satisfies d(ι(y),ι(x))=d(y,x)=d(x,y)<εd(\iota(y),\iota(x))=d(y,x)=d(x,y)<\varepsilon, by Metric Space (iv). So ι\iota is continuous at xx relative to AA by Continuous Map Between Metric Spaces (i); as x∈Ax\in A was arbitrary, ι\iota is continuous on AA by Continuous Map Between Metric Spaces (ii).

Proof of (ii). Let ff be continuous at x∈Ax\in A relative to AA and gg continuous at f(x)f(x) relative to BB, and let ε>0\varepsilon>0 be real. The choices are made in this order. First, Continuous Map Between Metric Spaces (i) for gg at f(x)f(x) gives a real η>0\eta>0 such that every s∈Bs\in B with dY(f(x),s)<ηd_{Y}(f(x),s)<\eta satisfies dZ(g(s),g(f(x)))<εd_{Z}(g(s),g(f(x)))<\varepsilon. Second, the same clause for ff at xx, applied to η\eta, gives a real Γ>0\delta>0 such that every u∈Au\in A with dX(x,u)<Γd_{X}(x,u)<\delta satisfies dY(f(u),f(x))<ηd_{Y}(f(u),f(x))<\eta. Let u∈Au\in A with dX(x,u)<Γd_{X}(x,u)<\delta. Then s=f(u)s=f(u) lies in BB and dY(f(x),s)=dY(f(u),f(x))<ηd_{Y}(f(x),s)=d_{Y}(f(u),f(x))<\eta, by Metric Space (iv); so

dZ((g∘f)(u),(g∘f)(x))=dZ(g(f(u)),g(f(x)))<ε.d_{Z}\bigl((g\circ f)(u),(g\circ f)(x)\bigr)=d_{Z}\bigl(g(f(u)),g(f(x))\bigr)<\varepsilon .

Hence g∘fg\circ f is continuous at xx relative to AA, by Continuous Map Between Metric Spaces (i). If ff is continuous on AA and gg on BB, then for every x∈Ax\in A, ff is continuous at xx relative to AA and gg at f(x)∈Bf(x)\in B relative to BB, by Continuous Map Between Metric Spaces (ii); so g∘fg\circ f is continuous at every x∈Ax\in A relative to AA, that is, continuous on AA.

Proof of (iii). Put r=1/gr=1/g. Continuity relative to EE is used in the form of Continuity of Real Functions on Subsets of ā„ in ε-Ī“ Form (i). For t∈Et\in E, g(t)≠0g(t)\neq0, so r(t)=1/g(t)=g(t)āˆ’1r(t)=1/g(t)=g(t)^{-1} and g(t) r(t)=1g(t)\,r(t)=1 (Difference and Quotient Identities in ā„ (ii)), and ∣r(t)∣=∣g(t)āˆ£āˆ’1|r(t)|=|g(t)|^{-1} (Difference and Quotient Identities in ā„ (iii)). For t∈Et\in E,

(†)r(t)āˆ’r(x)=(g(x)āˆ’g(t)) r(t) r(x),(\dagger)\qquad r(t)-r(x)=\bigl(g(x)-g(t)\bigr)\,r(t)\,r(x),

since the right side is g(x) r(x) r(t)āˆ’g(t) r(t) r(x)=r(t)āˆ’r(x)g(x)\,r(x)\,r(t)-g(t)\,r(t)\,r(x)=r(t)-r(x), by Rules of Arithmetic in a Commutative Ring: Zero, Signs and Squares, and No Zero Divisors in a Field (ii) for (uāˆ’v) w=u wāˆ’v w(u-v)\,w=u\,w-v\,w and by g(x) r(x)=1=g(t) r(t)g(x)\,r(x)=1=g(t)\,r(t).

Let gg be continuous at xx relative to EE. Put a=∣g(x)∣a=|g(x)|; then 0<a0<a, as 0≤a0\le a and a≠0a\neq0 (Rules of Arithmetic and Order in an Ordered Field (x)), and 0<a/20<a/2 and a/2+a/2=aa/2+a/2=a by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (vii). Put K=(a/2)āˆ’1 aāˆ’1K=(a/2)^{-1}\,a^{-1}; then 0<K0<K by Rules of Arithmetic and Order in an Ordered Field (viii) and Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (iv). Let ε>0\varepsilon>0 be real; then 0<ε Kāˆ’10<\varepsilon\,K^{-1} by the same two clauses. The choices are made in this order. First, continuity of gg at xx, applied to a/2a/2, gives a real Ī“1>0\delta_{1}>0 such that every t∈Et\in E with ∣tāˆ’x∣<Ī“1|t-x|<\delta_{1} satisfies ∣g(t)āˆ’g(x)∣<a/2|g(t)-g(x)|<a/2. Second, the same continuity, applied to ε Kāˆ’1\varepsilon\,K^{-1}, gives a real Ī“2>0\delta_{2}>0 such that every t∈Et\in E with ∣tāˆ’x∣<Ī“2|t-x|<\delta_{2} satisfies ∣g(t)āˆ’g(x)∣<ε Kāˆ’1|g(t)-g(x)|<\varepsilon\,K^{-1}. Third, take a real Ī“\delta with 0<Ī“0<\delta, Γ≤Γ1\delta\le\delta_{1} and Γ≤Γ2\delta\le\delta_{2}, such as min⁔{Ī“1,Ī“2}\min\{\delta_{1},\delta_{2}\}.

Let t∈Et\in E with ∣tāˆ’x∣<Ī“|t-x|<\delta, and put z=∣g(t)āˆ’g(x)∣z=|g(t)-g(x)|; then ∣tāˆ’x∣<Ī“1|t-x|<\delta_{1} and ∣tāˆ’x∣<Ī“2|t-x|<\delta_{2}, by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (i). Also ∣g(x)āˆ’g(t)∣=z|g(x)-g(t)|=z, as g(x)āˆ’g(t)=āˆ’(g(t)āˆ’g(x))g(x)-g(t)=-(g(t)-g(x)) (Difference and Quotient Identities in ā„ (i)) and āˆ£āˆ’w∣=∣w∣|-w|=|w| (Rules of Arithmetic and Order in an Ordered Field (x)). If z=0z=0, then g(t)āˆ’g(x)=0g(t)-g(x)=0 by Rules of Arithmetic and Order in an Ordered Field (x), so g(t)=g(x)g(t)=g(x), r(t)=r(x)r(t)=r(x) and ∣r(t)āˆ’r(x)∣=∣0∣=0<ε|r(t)-r(x)|=|0|=0<\varepsilon. Let now z≠0z\neq0, so that 0<z0<z. By Rules of Arithmetic and Order in an Ordered Field (xi) and the first choice,

a=∣(g(x)āˆ’g(t))+g(t)āˆ£ā‰¤z+∣g(t)∣<a/2+∣g(t)∣,a=\bigl|\bigl(g(x)-g(t)\bigr)+g(t)\bigr|\le z+|g(t)|<a/2+|g(t)|,

the strict inequality by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (ii); so a/2+a/2=a<a/2+∣g(t)∣a/2+a/2=a<a/2+|g(t)| by Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (i), and a/2<∣g(t)∣a/2<|g(t)| by Rules of Arithmetic and Order in an Ordered Field (iv). Hence ∣r(t)∣=∣g(t)āˆ£āˆ’1<(a/2)āˆ’1|r(t)|=|g(t)|^{-1}<(a/2)^{-1} by Rules of Arithmetic and Order in an Ordered Field (viii). By (†)(\dagger), Rules of Arithmetic and Order in an Ordered Field (x) and ∣r(x)∣=aāˆ’1|r(x)|=a^{-1},

∣r(t)āˆ’r(x)∣=zā€‰āˆ£r(t)āˆ£ā€‰aāˆ’1<z (a/2)āˆ’1 aāˆ’1=z K<ε Kāˆ’1 K=ε,|r(t)-r(x)|=z\,|r(t)|\,a^{-1}<z\,(a/2)^{-1}\,a^{-1}=z\,K<\varepsilon\,K^{-1}\,K=\varepsilon ,

where the first inequality multiplies ∣r(t)∣<(a/2)āˆ’1|r(t)|<(a/2)^{-1} by z aāˆ’1z\,a^{-1}, positive by Rules of Arithmetic and Order in an Ordered Field (viii) and Inequalities in an Ordered Field: Mixed Transitivity, Strict Sums, Signs, Products, Natural Numbers, Halving, Reciprocals, Absolute Values and Squares (iv), and the second multiplies z<ε Kāˆ’1z<\varepsilon\,K^{-1}, the second choice, by K>0K>0, both by Rules of Arithmetic and Order in an Ordered Field (vi). So rr is continuous at xx relative to EE, by Continuity of Real Functions on Subsets of ā„ in ε-Ī“ Form (i).

If gg is continuous on EE, then it is continuous at every x∈Ex\in E relative to EE (Continuous Map Between Metric Spaces (ii)); by the above, so is rr, which is therefore continuous on EE, by the same clause.

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…