Processing math: 3%

B.1 Limit Theorems and Proofs

Printed Page B-1

Uniqueness of a Limit

The limit of a function f, if it exists, is unique; that is, a function can only have one limit.

THEOREM A Limit Is Unique

If a function f is defined on an open interval containing the number c, except possibly at c itself, and if lim and \lim\limits_{x\rightarrow c}f(x) = L_{2}, then L_{1} = L_{2}.

Recall

An indirect proof (a proof by contradiction) begins by assuming the conclusion is false. Then we show that this assumption leads to a contradiction.

Proof

Assume that L_{1}\neq L_{2}. We will show that this assumption leads to a contradiction. By the definition of the limit of a function f, \lim\limits_{x\rightarrow c}f(x) = L_{1} if, for any given number \varepsilon > 0, there is a number \delta _{1} > 0, so that \begin{equation} {\rm whenever}\ 0 < \vert x-c\vert < \delta_{1}\quad {\rm then}\ \vert f(x)-L_{1}\vert < \varepsilon\tag{1} \end{equation}

Recall

The Triangle Inequality: If {x} and {y} are real numbers, then \vert {x+y}\vert \leq \vert{x}\vert + \vert{y}\vert. See Appendix A.1, p. A-7.

Similarly, \lim\limits_{x\rightarrow c}f(x) = L_{2} if, for any given number \varepsilon > 0, there is a number \delta_{2} > 0, so that \begin{equation} {\rm whenever}\ 0 < \vert x-c\vert < \delta_{2}\quad {\rm then }\ \vert f(x)-L_{2}\vert < \varepsilon\tag{2} \end{equation}

Now \begin{equation*} L_{1} - L_{2} = L_{1}-f(x) + f(x)-L_{2} \end{equation*}

so, by applying the Triangle Inequality, we get \begin{equation} \vert L_{1}-L_{2}\vert = \vert L_{1}-f(x)+f(x)-L_{2}\vert \leq \vert L_{1}-f(x)\vert + \vert f(x)-L_{2}\vert\tag{3} \end{equation}

For any given number \varepsilon > 0, let \delta be the smaller of \delta _{1} and \delta _{2}. Then from (1)–(3), we can conclude that whenever 0 < \vert x-c \vert < \delta \leq \delta_{1} and 0 < \vert x-c\vert < \delta \leq \delta _{2}, we have \begin{equation} \vert L_{1}-L_{2}\vert < \varepsilon +\varepsilon = 2 \varepsilon \tag{4} \end{equation}

In particular, (4) is true for \varepsilon =\dfrac{1}{2}\vert L_{1}-L_{2}\vert >0. (Remember L_{1}\neq L_{2}.) Then from (4), \begin{equation*} \vert L_{1}-L_{2}\vert < 2\varepsilon = \vert L_{1}-L_{2}\vert \end{equation*}

which is a contradiction. Therefore, L_{1}=L_{2}, and the limit, if it exists, is unique.

Algebra of Limits

THEOREM Limit of a Sum

If f and g are functions for which \lim\limits_{x\rightarrow c}f( x) and \lim\limits_{x\rightarrow c}g(x) both exist, then \lim\limits_{x\rightarrow c}\,[f(x)+g(x)] exists and \lim\limits_{x\rightarrow c}[f(x)+g(x)]=\lim\limits_{x\rightarrow c}f(x)+\lim\limits_{x\rightarrow c}g(x)

B-2

Proof

Suppose \lim\limits_{x\rightarrow c}f(x)=L and \lim\limits_{x\rightarrow c}g(x)=M. We need to show that for any number \varepsilon > 0, there is a number \delta > 0, so that \begin{equation*} {\rm whenever}\ 0 < \vert x-c\vert < \delta\quad {\rm then}\ \vert [ f(x)+g(x)] - [ L+M] \vert < \varepsilon \end{equation*}

Since \lim\limits_{x\rightarrow c}f(x) =L, by the definition of a limit, given the number \dfrac{\varepsilon }{2} > 0, there is a number \delta_{1} > 0, so that \begin{equation*} {\rm whenever}\ 0 < \vert x-c\vert < \delta_{1}\quad {\rm then}\ \vert f(x)-L\vert < \dfrac{\varepsilon }{2} \end{equation*}

Since \lim\limits_{x\rightarrow c}g(x)=M, for this same number \dfrac{\varepsilon }{2}, there is a number \delta _{2} > 0, so that \begin{equation*} {\rm whenever}\ 0 < \vert x-c\vert < \delta _{2}\quad {\rm then}\ \vert g(x)-M\vert < \dfrac{\varepsilon }{2} \end{equation*}

Let \delta be the smaller of \delta _{1} and \delta _{2}. Then \delta \leq \delta _{1} and \delta \leq \delta _{2}. Using this \delta, \begin{eqnarray*} {\rm whenever}\ 0 & < \vert x-c\vert <\delta\quad {\rm then}\ \vert f(x)-L\vert < \dfrac{\varepsilon }{2} \\[4.5pt] {\rm whenever}\ 0 & < \vert x-c\vert < \delta\quad {\rm then}\ \vert g(x)-M\vert < \dfrac{\varepsilon }{2} \end{eqnarray*}

That is, whenever 0 < \vert x-c\vert < \delta, \begin{eqnarray*} \begin{array}{ll} \vert [ f(x)+g(x)] -[ L+M]\vert & = \vert [ f(x)-L] + [ g(x)-M] \vert \\[4pt] & \leq \vert f(x)-L\vert + \vert g(x)-M \vert\quad {\color{#0066A7}{\hbox{Use the Triangle Inequality.}}} \\[4pt] & < \dfrac{\varepsilon}{2} + \dfrac{\varepsilon }{2} = \varepsilon \end{array}\end{eqnarray*}So, \lim\limits_{x\rightarrow c} [ f(x)+g(x)] =L+M.

THEOREM Limit of a Product

If f and g are functions for which \lim\limits_{x\rightarrow c}f(x) and \lim\limits_{x\rightarrow c}g(x) both exist,
then \lim\limits_{x\rightarrow c}\,[f(x)\cdot g(x)] exists and \begin{eqnarray*} \lim\limits_{x\rightarrow c}\,[f(x)\cdot g(x)]=\lim\limits_{x\rightarrow c}f(x)\cdot \lim\limits_{x\rightarrow c}g(x) \end{eqnarray*}

Proof

Suppose \lim\limits_{x\rightarrow c}f(x) = L and \lim\limits_{x\rightarrow c}g(x) = M. We need to show that for any number \varepsilon > 0, there is a number \delta > 0, so that \begin{equation*} {\rm whenever}\ 0 < \vert x-c\vert < \delta\quad {\rm then}\ \vert f(x)\cdot g(x)-L\cdot M\vert < \varepsilon \end{equation*}

Subtracting and adding f(x)\cdot M in the expression f(x)\cdot g(x)-L\cdot M result in terms involving g(x)-M and f(x)-L:

Since \lim\limits_{x\rightarrow c}f(x)=L and \lim\limits_{x\rightarrow c}g(x)=M, then there is a number \delta _{1} > 0, so that whenever 0 < \vert x-c\vert <\delta _{1}, then \begin{equation} \vert f(x)-L\vert < 1, \qquad \hbox{from which}\qquad \vert f(x)\vert < 1 + \vert L\vert\tag{6} \end{equation}

Also given a number \varepsilon >0, there is a number \delta _{2}, so that whenever 0 < \vert x-c\vert < \delta_{2}, then \begin{equation} \vert g(x)-M\vert < \dfrac{\varepsilon }{1+ \vert L\vert +\vert M\vert }\tag{7} \end{equation}

B-3

Given a number \varepsilon >0, there is a number \delta _{3}, so that whenever 0 < \vert x-c \vert <\delta _{3}, then \begin{equation} \vert f(x)-L\vert <\dfrac{\varepsilon }{1+\vert L\vert +\vert M\vert }\tag{8} \end{equation}

Choose \delta to be the minimum of \delta _{1}, \delta _{2}, and \delta _{3} and combine (5)–(8). Then for any given \varepsilon >0, there is a \delta >0, so that whenever 0 < \vert x-c\vert <\delta, we have \begin{eqnarray*} \vert f(x)\cdot g(x)-L\cdot M\vert & < & [ 1 + \vert L\vert ] \dfrac{\varepsilon }{1 + \vert L\vert + \vert M\vert } + \vert M\vert \dfrac{\varepsilon }{ 1 + \vert L\vert + \vert M\vert } \\[4pt] & < & [ 1 + \vert L\vert + \vert M\vert ] \dfrac{ \varepsilon }{1 + \vert L\vert + \vert M\vert } =\varepsilon \end{eqnarray*}

That is, \lim\limits_{x\ c}\,[f(x)\cdot g(x)]=\lim\limits_{x\rightarrow c}f(x)\cdot \lim\limits_{x\rightarrow c}g(x).

THEOREM Squeeze Theorem

If the functions f, g, and h have the property that for all x in an open interval containing c, except possibly at c itself, f(x)\leq g(x)\leq h(x) and if \lim_{x\rightarrow c}f(x) = \lim_{x\rightarrow c}h(x) = L then \lim_{x\rightarrow c}g(x) = L

Proof

Since \lim\limits_{x\rightarrow c}f(x) = \lim\limits_{x\rightarrow c}h(x) = L, then for any number \varepsilon > 0, there are positive numbers \delta _{1} and \delta _{2}, so that \begin{eqnarray*} {\rm whenever}\ 0 & < \vert x-c \vert < \delta_{1}\quad \hbox{then } \vert f(x)-L \vert < \varepsilon\\ {\rm whenever}\ 0 & < \vert x-c\vert < \delta_{2}\quad \hbox{then } \vert h(x)-L \vert < \varepsilon \end{eqnarray*}

Choose \delta to be the smaller of the numbers \delta_{1} and \delta_{2}. Then 0 < \vert x-c\vert < \delta implies that both \vert f(x)-L \vert < \varepsilon and \vert h(x)-L \vert < \varepsilon. In other words, 0 < \vert x-c \vert < \delta implies that both \begin{equation*} L - \varepsilon < f(x) < L + \varepsilon\qquad \hbox{and}\qquad L - \varepsilon < h(x) < L + \varepsilon \end{equation*}

Since f(x)\leq g(x)\leq h(x) for all x\neq c in the open interval, it follows that whenever 0 < \vert x-c\vert < \delta and x is in the open interval, we have \begin{equation*} L - \varepsilon < f(x) \leq g(x) \leq h(x) < L + \varepsilon \end{equation*}

Then for any given number \varepsilon >0, there is a positive number \delta , so that whenever 0 < \vert x-c\vert < \delta, then L - \varepsilon < g(x) < L + \varepsilon, or equivalently, \vert g(x) -L\vert < \varepsilon. That is, \lim\limits_{x\rightarrow c}g(x) = L.