5.1: Limits
( \newcommand{\kernel}{\mathrm{null}\,}\)
Let A⊂R and let x be a limit point of A. In the following, we will let S(A,x) denote the set of all convergent sequences {xn}n∈I such that xn∈A for all n∈I,xn≠x for all n∈I, and limn→∞xn=x. We will let S+(A,x) be the subset of S(A,x) of sequences {xn}n∈I for which xn>x for all n∈I and S−(A,x) be the subset of S(A,x) of sequences {xn}n∈I for which xn<x for all n∈I.
Let D⊂R,f:D→R,L∈R, and suppose a is a limit point of D. We say the limit of f as x approaches a is L, denoted
limx→af(x)=L,
if for every sequence {xn}n∈I∈S(D,a),
limn→∞f(xn)=L.
If S+(D,a)≠∅, we say the limit from the right of f as x approaches a is L, denoted
limx→a+f(x)=L,
if for every sequence {xn}n∈I∈S+(D,a),
limn→∞f(xn)=L,
and, if S−(D,a)≠∅, we say the limit from the left of f as x approaches a is L, denoted
limx→a−f(x)=L,
if for every sequence {xn}n∈I∈S−(D,a),
limn→∞f(xn)=L.
We may also denote
limx→af(x)=L
by writing
f(x)→L as x→a.
Similarly, we may denote
limx→a+f(x)=L
by writing
f(x)→L as x↓a
and
limx→a−f(x)=L
by writing
f(x)→L as x↑a
We also let
f(a+)=limx→a+f(x)
and
f(a−)=limx→a−f(x).
It should be clear that if limx→af(x)=L and S+(D,a)≠∅, then f(a+)=L. Similarly, if limx→af(x)=L and S−(D,a)≠∅, then f(a−)=L.
Suppose D⊂R,f:D→R, and a is a limit point of D. If f(a−)=f(a+)=L, then limx→af(x)=L.
- Proof
-
Suppose {xn}∞n=m∈S(D,a). Let
J−={n:n∈Z,xn<a}
and
J+={n:n∈Z,xn>a}.
Suppose J− is empty or finite and let k=m−1 if J−=∅ and, otherwise, let k be the largest integer in J−. Then {xn}∞n=k+1∈S+(D,a), and so
limn→∞f(xn)=f(a+)=L.
A similar argument shows that if J+ is empty or finite, then
limn→∞f(xn)=f(a−)=L.
If neither J− nor J+ is finite or empty, then {xn}n∈J− and {xn}n∈J+ are subsequences of {xn}∞n=m with {xn}n∈J−∈S−(D,a) and {xn}n∈J+∈S+(D,a). Hence, given any ϵ>0, we may find integers N and M such that
|f(xn)−L|<ϵ
whenever n∈{j:j∈J−,j>N} and
|f(xn)−L|<ϵ
whenever n∈{j:j∈J+,j>M}. Let P be the larger of N and M. Since J−∪J+={j:j∈Z+,j≥m}, it follows that
|f(xn)−L|<ϵ
whenever n>P. Hence limn→∞f(xn)=L, and so limx→af(x)=L. Q.E.D.
Suppose D⊂R,a is a limit point of D, and f:D→R. If limx→af(x)=L and α∈R, then
limx→aαf(x)=αL.
- Proof
-
Suppose {xn}n∈I∈S(D,a). Then
limn→∞αf(xn)=αlimn→∞f(xn)=αL.
Hence limx→aαf(x)=αL. Q.E.D.
Suppose D⊂R,a is a limit point of D,f:D→R, and g:D→R. If limx→af(x)=L and limx→ag(x)=M, then
limx→a(f(x)+g(x))=L+M.
- Proof
-
Suppose {xn}n∈I∈S(D,a). Then
limn→∞(f(xn)+g(xn))=limn→∞f(xn)+limn→∞g(xn)=L+M.
Hence limx→a(f(x)+g(x))=L+M. Q.E.D.
Suppose D⊂R,a is a limit point of D,f:D→R, and g:D→R. If limx→af(x)=L and limx→ag(x)=M, then
limx→af(x)g(x)=LM.
Prove the previous proposition.
Suppose D⊂R,a is a limit point of D,f:D→R, g:D→R, and g(x)≠0 for all x∈D. If limx→af(x)=L,limx→ag(x)=M, and M≠0, then
limx→af(x)g(x)=LM.
Prove the previous proposition.
Suppose D⊂R,a is a limit point of D,f:D→R, and f(x)≥0 for all x∈D. If limx→af(x)=L, then
limx→a√f(x)=√L.
Prove the previous proposition.
Given D⊂R,f:D→R, and A⊂D, we let
f(A)={y:y=f(x) for some x∈A}.
In particular, f(D) denotes the range of f.
Suppose D⊂R,E⊂R,a is a limit point of D,g:D→R, f:E→R, and g(D)⊂E. Moreover, suppose limx→ag(x)=b and, for some ϵ>0, g(x)≠b for all x∈(a−ϵ,a+ϵ)∩D. If limx→bf(x)=L, then
limx→af∘g(x)=L.
- Proof
-
Suppose {xn}n∈I∈S(D,a). Then
limn→∞g(xn)=b.
Let N∈Z+ such that |xn−a|<ϵ whenever n>N. Then
{g(xn)}∞n=N+1∈S(E,b),
so
limn→∞f(g(xn))=L.
Thus limx→af∘g(x)=L. Q.E.D.
Let
g(x)={0, if x≠0,1, if x=0.
If f(x)=g(x), then
f∘g(x)={1, if x≠0,0, if x=0.
Hence limx→0f∘g(x)=1, although limx→0g(x)=0 and limx→0f(x)=0.
5.1.1 Limits of Polynomials and Rational Functions
If c∈R and f:R→R is given by f(x)=c for all x∈R, then clearly limx→af(x)=c for any a∈R.
Suppose f:R→R is defined by f(x)=x for all x∈R. If, for any a∈R,{xn}n∈I∈S(R,a), then
limn→∞f(xn)=limn→∞xn=a.
Hence limx→ax=a.
Suppose n∈Z+ and f:R→R is defined by f(x)=xn. Then
limx→af(x)=limx→axn=n∏i=1limx→ax=an.
If n∈Z,n≥0, and b_{0}, b_{1}, \ldots, b_{n} are real numbers with b_{n} \neq 0, then we call the function p: \mathbb{R} \rightarrow \mathbb{R} defined by
p(x)=b_{n} x^{n}+b_{n-1} x^{n-1}+\cdots+b_{1} x+b_{0}
a polynomial of degree n.
Show that if f is a polynomial and a \in \mathbb{R}, then \lim _{x \rightarrow a} f(x)=f(a).
Suppose p and q are polynomials and
D=\{x: x \in \mathbb{R}, q(x) \neq 0\}.
We call the function r: D \rightarrow \mathbb{R} defined by
r(x)=\frac{p(x)}{q(x)}
a rational function.
Show that if f is a rational function and a is in the domain of f, then \lim _{x \rightarrow a} f(x)=f(a).
Suppose D \subset \mathbb{R}, a \in D is a limit point of D, and \lim _{x \rightarrow a} f(x)=L. If E=D \backslash\{a\} and g: E \rightarrow \mathbb{R} is defined by g(x)=f(x) for all x \in E, show that \lim _{x \rightarrow a} g(x)=L .
Evaluate
\lim _{x \rightarrow 1} \frac{x^{5}-1}{x^{3}-1}.
Suppose D \subset \mathbb{R}, a is a limit point of D, f: D \rightarrow \mathbb{R}, g: D \rightarrow \mathbb{R}, h: D \rightarrow \mathbb{R}, and f(x) \leq h(x) \leq g(x) for all x \in D . If \lim _{x \rightarrow a} f(x)=L and \lim _{x \rightarrow a} g(x)=L, show that \lim _{x \rightarrow a} h(x)=L . (This is the squeeze theorem for limits of functions.)
Note that the above results which have been stated for limits will hold as well for the appropriate one-sided limits, that is, limits from the right or from the left.
Suppose
f(x)=\left\{\begin{array}{ll}{x+1,} & {\text { if } x<0,} \\ {4,} & {\text { if } x=0,} \\ {x^{2},} & {\text { if } x>0.}\end{array}\right.
Evaluate f(0), f(0-), and f(0+) . Does \lim _{x \rightarrow 0} f(x) exist?
5.1.2 Equivalent Definitions
Suppose D \subset \mathbb{R}, a is a limit point of D, and f: D \rightarrow \mathbb{R}. Then \lim _{x \rightarrow a} f(x)=L if and only if for every \epsilon>0 there exists a \delta>0 such that
|f(x)-L|<\epsilon \text { whenever } x \neq a \text { and } x \in(a-\delta, a+\delta) \cap D.
- Proof
-
Suppose \lim _{x \rightarrow a} f(x)=L . Suppose there exists an \epsilon>0 such that for every \delta>0 there exists x \in(a-\delta, a+\delta) \cap D, x \neq a, for which |f(x)-L| \geq \epsilon. For n=1,2,3, \ldots, choose
x_{n} \in\left(a-\frac{1}{n}, a+\frac{1}{n}\right) \cap D,
x_{n} \neq a, such that \left|f\left(x_{n}\right)-L\right| \geq \epsilon . Then \left\{x_{n}\right\}_{n=1}^{\infty} \in S(D, a), but \left\{f\left(x_{n}\right)\right\}_{n=1}^{\infty} does not converge to L, contradicting the assumption that \lim _{x \rightarrow a} f(x)=L.
Now suppose that for every \epsilon>0 there exists \delta>0 such that |f(x)-L|<\epsilon whenever x \neq a and x \in(a-\delta, a+\delta) \cap D . Let \left\{x_{n}\right\}_{n \in I} \in S(D, a) . Given \epsilon>0, let \delta>0 be such that |f(x)-L|<\epsilon whenever x \neq a and x \in(a-\delta, a+\delta) \cap D . Choose N \in \mathbb{Z} such that \left|x_{n}-a\right|<\delta whenever n>N . Then \left|f\left(x_{n}\right)-L\right|<\epsilon for all n>N . Hence \lim _{n \rightarrow \infty} f\left(x_{n}\right)=L, and so \lim _{x \rightarrow a} f(x)=L . \quad Q.E.D.
The proofs of the next two propositions are analogous.
Suppose D \subset \mathbb{R}, a is a limit point of D, f: D \rightarrow \mathbb{R}, and S^{-}(D, a) \neq \emptyset . Then \lim _{x \rightarrow a^{-}} f(x)=L if and only if for every \epsilon>0 there exists a \delta>0 such that
|f(x)-L|<\epsilon \text { whenever } x \in(a-\delta, a) \cap D.
Suppose D \subset \mathbb{R}, a is a limit point of D, f: D \rightarrow \mathbb{R}, and S^{+}(D, a) \neq \emptyset . Then \lim _{x \rightarrow a^{+}} f(x)=L if and only if for every \epsilon>0 there exists a \delta>0 such that
|f(x)-L|<\epsilon \text { whenever } x \in(a, a+\delta) \cap D.
5.1.3 Examples
Define f: \mathbb{R} \rightarrow \mathbb{R} by
f(x)=\left\{\begin{array}{ll}{1,} & {\text { if } x \text { is rational, }} \\ {0,} & {\text { if } x \text { is irrational. }}\end{array}\right.
Let a \in \mathbb{R} . Since every open interval contains both rational and irrational numbers, for any \delta>0 and any choice of L \in \mathbb{R}, there will exist x \in(a-\delta, a+\delta), x \neq a, such that
|f(x)-L| \geq \frac{1}{2}.
Hence \lim _{x \rightarrow a} f(x) does not exist for any real number a.
Define f: \mathbb{R} \rightarrow \mathbb{R} by
f(x)=\left\{\begin{array}{ll}{x,} & {\text { if } x \text { is rational, }} \\ {0,} & {\text { if } x \text { is irrational. }}\end{array}\right.
Then \lim _{x \rightarrow 0} f(x)=0 since, given \epsilon>0,|f(x)|<\epsilon provided |x|<\epsilon.
Show that if f is as given in the previous example and a \neq 0, then \lim _{x \rightarrow a} f(x) does not exist.
Define f: \mathbb{R} \rightarrow \mathbb{R} by
f(x)=\left\{\begin{array}{ll}{\frac{1}{q},} & {\text { if } x \text { is rational and } x=\frac{p}{q},} \\ {0,} & {\text { if } x \text { is irrational, }}\end{array}\right.
where p and q are taken to be relatively prime integers with q>0, and we take q=1 when x=0 . Show that, for any real number a, \lim _{x \rightarrow a} f(x)=0.
Define \varphi:[0,1] \rightarrow[-1,1] by
\varphi(x)=\left\{\begin{array}{ll}{4 x,} & {\text { if } 0 \leq x \leq \frac{1}{4},} \\ {2-4 x,} & {\text { if } \frac{1}{4}<x<\frac{3}{4},} \\ {4 x-4,} & {\text { if } \frac{3}{4} \leq x \leq 1.}\end{array}\right.
Next define s: \mathbb{R} \rightarrow \mathbb{R} by s(x)=\varphi(x-\lfloor x\rfloor), where \lfloor x\rfloor denotes the largest integer less than or equal to x (that is, \lfloor x\rfloor is the floor of x ). The function s is an example of a sawtooth function. See the graphs of \varphi and s in Figure 5.1 .1 . Note that for any n \in \mathbb{Z},
s([n, n+1])=[-1,1].
Now let D=\mathbb{R} \backslash\{0\} and define \sigma: D \rightarrow \mathbb{R} by
\sigma(x)=s\left(\frac{1}{x}\right).
See the graph of \sigma in Figure 5.1 .2 . Note that for any n \in \mathbb{Z}^{+},
\sigma\left(\left[\frac{1}{n+1}, \frac{1}{n}\right]\right)=s([n, n+1])=[-1,1].
Hence for any \epsilon>0, \sigma((0, \epsilon))=[-1,1], and so \lim _{x \rightarrow 0^{+}} \sigma(x) does not exist. Similarly, neither \lim _{x \rightarrow 0^{-}} \sigma(x) nor \lim _{x \rightarrow 0} \sigma(x) exist.
Let s be the sawtooth function of the previous example and let D=\mathbb{R} \backslash\{0\} . Define \psi: D \rightarrow \mathbb{R} by
\psi(x)=x s\left(\frac{1}{x}\right).
See Figure 5.1 .2 for the graph of \psi . Then for all x \in D,
-|x| \leq \psi(x) \leq|x|,
and so \lim _{x \rightarrow 0} \psi(x)=0 by the squeeze theorem.
Let D \subset \mathbb{R} and f: D \rightarrow \mathbb{R} . We say f is bounded if there exists a real number B such that |f(x)| \leq B for all x \in D.
Suppose f: \mathbb{R} \rightarrow \mathbb{R} is bounded. Show that \lim _{x \rightarrow 0} x f(x)=0.