next up previous index
Next: 9. Trigonometric Functions Up: 8. Integrable Functions Previous: 8.5 Change of Scale   Index

8.6 Integrals and Area

8.51   Theorem. Let $f$ be a piecewise monotonic function from an interval $[a,b]$ to $\mbox{{\bf R}}_{\geq 0}$. Then

\begin{displaymath}\int_a^b f=A_a^b f=\alpha(S_a^b f).\end{displaymath}

Proof: We already know this result for monotonic functions, and from this the result follows easily for piecewise monotonic functions. $\diamondsuit$


Remark Theorem 8.51 is in fact true for all integrable functions from $[a,b]$ to $\mbox{{\bf R}}_{\geq 0}$, but the proof is rather technical. Since we will never need the result for functions that are not piecewise monotonic, I will not bother to make an assumption out of it.

8.52   Theorem. Let $a,b\in\mbox{{\bf R}}$ and let $f\colon
[a,b]\to\mbox{{\bf R}}$ be a piecewise monotonic function. Then the graph of $f$ is a zero-area set.

Proof: We will show that the theorem holds when $f$ is monotonic on $[a,b]$. It then follows easily that the theorem holds when $f$ is piecewise monotonic on $[a,b]$.

Suppose $f$ is increasing on $[a,b]$. Let $n\in\mbox{${\mbox{{\bf Z}}}^{+}$}$ and let $P=\{x_0,x_1,\cdots ,x_n\}$ be the regular partition of $[a,b]$ into $n$ equal subintervals.

\psfig{file=ch8j.eps,width=1.25in}
Then

\begin{displaymath}x_i-x_{i-1}={{b-a}\over n} \mbox{ for } 1\leq i\leq n\end{displaymath}

and

\begin{displaymath}\mbox{graph}(f)\subset\bigcup_{i=1}^n B\Big(x_{i-1},x_i\colon
f(x_{i-1}),f(x_i)\Big).\end{displaymath}

Hence

\begin{eqnarray*}
0\leq\alpha\Big(\mbox{graph}(f)\Big)
&\leq& \alpha \Big( \big...
...(x_i)-f(x_{i-1})\Big)\cr
&=& {{b-a}\over n}\Big( f(b)-f(a)\Big).
\end{eqnarray*}



Now $\displaystyle {\Big\{ {{b-a}\over n}\Big( f(b)-f(a)\Big)\Big\} \to 0}$, so it follows from the squeezing rule that the constant sequence $\Big\{ \alpha \Big(\mbox{graph}(f)\Big)\Big\}$ converges to $0$, and hence

\begin{displaymath}\alpha\Big(\mbox{graph}(f)\Big)=0.\mbox{ $\diamondsuit$}\end{displaymath}

Remark: Theorem 8.52 is actually valid for all integrable functions on $[a,b]$.

8.53   Theorem (Area between graphs.) Let $f,g$ be piecewise monotonic functions on an interval $[a,b]$ such that $g(x)\leq
f(x)$ for all $x\in [a,b]$. Let

\begin{displaymath}S=\{(x,y)\colon a\leq x\leq b \mbox{ {\rm and} } g(x)\leq y\leq f(x)\}.\end{displaymath}

Then

\begin{displaymath}\mbox{{\rm area}}(S)=\int_a^b f(x)-g(x)\; dx.\end{displaymath}


Proof: Let $M$ be a lower bound for $g$, so that

\begin{displaymath}0\leq g(x)-M\leq f(x)-M \mbox{ for all } x\in [a,b].\end{displaymath}

Let

\begin{displaymath}F(x)=f(x)-M,\qquad G(x)=g(x)-M\end{displaymath}

for all $x\in [a,b]$, and let

\begin{displaymath}T=\{(x,y)\colon a\leq x\leq b \mbox{ and } G(x)\leq y\leq F(x)\}.\end{displaymath}

\psfig{file=ch8k.eps,width=1.8in} \psfig{file=ch8l.eps,width=1.8in}
Then

\begin{eqnarray*}
(x,y)\in T &\iff& a\leq x\leq b \mbox{ and } G(x)\leq y\leq F(...
... y+M\leq f(x)\\
&\iff& (x,y+M)\in S\\
&\iff& (x,y)+(0,M)\in S.
\end{eqnarray*}



It follows from translation invariance of area that

\begin{displaymath}\mbox{area}(S)=\mbox{area}(T).\end{displaymath}

Let

\begin{eqnarray*}
R &=& \{ (x,y)\colon a\leq x\leq b \mbox{ and } 0\leq y\leq F(...
...y)\colon a\leq x\leq b \mbox{ and } 0\leq y\leq G(x)\} =S_a^b G.
\end{eqnarray*}



\psfig{file=ch8m.eps,width=1.8in} \psfig{file=ch8n.eps,width=1.8in}
Then $V\cup T=R$, and

\begin{displaymath}V\cap T=\{(x,y)\colon a\leq x\leq b \mbox{ and } y=G(x)\}=\mbox{graph}(G).\end{displaymath}

It follows from theorem 8.52 that $V$ and $T$ are almost disjoint, so

\begin{displaymath}\mbox{\rm area}(R)=\mbox{area}(V\cup
T)=\mbox{area}(V)+\mbox{area}(T),\end{displaymath}

and thus

\begin{displaymath}\mbox{\rm area}(T)=\mbox{area}(R)-\mbox{area}(V).\end{displaymath}

By theorem 8.51 we have

\begin{displaymath}\mbox{area}(R)=\mbox{area}(S_a^b F)=\int_a^b F(x)\; dx\end{displaymath}

and

\begin{displaymath}\mbox{area}(V)=\mbox{area}(S_a^b G)=\int_a^b G(x)\; dx.\end{displaymath}

Thus

\begin{eqnarray*}
\mbox{area}(S) &=& \mbox{\rm area}(T) = \mbox{\rm area}(R) - \...
...Bigg)\; dx \\
&=&\int_a^b f(x)-g(x)\; dx.\mbox{ $\diamondsuit$}
\end{eqnarray*}




Remark: Theorem 8.53 is valid for all integrable functions $f$ and $g$. This follows from our proof and the fact that theorems 8.51 and 8.52 are both valid for all integrable functions.

8.54   Example. We will find the area of the set $R$ in the figure, which is bounded by the graphs of $f$ and $g$ where

\begin{displaymath}f(x) = {1\over 2} x^2 \end{displaymath}

and

\begin{displaymath}g(x) = x^3 - 3x^2 + 3x. \end{displaymath}

\psfig{file=ch8o.eps,width=2in}
Now

\begin{eqnarray*}
g(x) - f(x) &=& x^3 - 3x^2 + 3x - {1\over 2}x^2 = x^3 - {7 \ov...
...+ 3x\\
&=& x(x^2 - {7\over 2} x + 3) = x(x-2)(x - {3\over 2}).
\end{eqnarray*}



Hence

\begin{displaymath}\big( g(x) = f(x) \big) \hspace{1ex}\Longleftrightarrow\hspac...
...space{1ex}\left( x \in \left\{0,{3\over 2},2 \right\} \right). \end{displaymath}

It follows that the points $\mbox{{\bf p}}$ and $\mbox{{\bf q}}$ in the figure are

\begin{displaymath}\mbox{{\bf p}}= ({3\over 2}, f({3\over 2})) = ({3\over 2},{9\over 8}) \mbox{ and }
\mbox{{\bf q}}= (2,f(2)) = (2,2). \end{displaymath}

Also, since $x(x-2) \leq 0$ for all $x \in [0,2]$,

\begin{displaymath}g(x) - f(x) \geq 0 \hspace{1ex}\Longleftrightarrow\hspace{1ex...
...0 \hspace{1ex}\Longleftrightarrow\hspace{1ex}x \leq {3\over 2}.\end{displaymath}

(This is clear from the picture, assuming that the picture is accurate.) Thus

\begin{eqnarray*}
\mbox{\rm area}(R) &=& \int_0^{3\over 2} (g - f) + \int_{3\ove...
...} \over 3\right)
-3\left({2^2 -{({3\over 2})}^2 \over 2}\right).
\end{eqnarray*}



We have now found the area, but the answer is not in a very informative form. It is not clear whether the number we have found is positive. It would be reasonable to use a calculator to simplify the result, but my experience with calculators is that I am more likely to make an error entering this into my calculator than I am to make an error by doing the calculation myself, so I will continue. I notice that three terms in the answer are repeated twice, so I have

\begin{eqnarray*}
\mbox{\rm area}(R) &=& 2\left( {81 \over 64} - {63 \over 16} +...
...{21 \over 32 } +{1\over 12} = {63 + 8 \over 96 } = {71\over 96}.
\end{eqnarray*}



Thus the area is about $0.7$ From the sketch I expect the area to be a little bit smaller than $1$, so the answer is plausible.

8.55   Exercise. The curve whose equation is
\begin{displaymath}
y^2+2xy+2x^2=4
\end{displaymath} (8.56)

is shown in the figure. Find the area enclosed by the curve.
\psfig{file=ch8p.eps,height=2.in}
(The set whose area we want to find is bounded by the graphs of the two functions. You can find the functions by considering equation (8.56) as a quadratic equation in $y$ and solving for $y$ as a function of $x$.) A

8.57   Exercise. A Find the areas of the two sets shaded in the figures below:
\psfig{file=ch8q.eps,width=5in}

8.58   Exercise. A Find the area of the shaded region.
\psfig{file=ch8s.eps,height=3in}


next up previous index
Next: 9. Trigonometric Functions Up: 8. Integrable Functions Previous: 8.5 Change of Scale   Index
Ray Mayer 2007-09-07