Skip to main content

Section 5.3 Exponential Functions Numerically, Modeling

How do we recognize that a function given numericallyβ€”by a table of valuesβ€”is an exponential function? This is an important question in applications. Typically, when you study a real-life process you do not have a formula for a function describing the process ahead of time. You take measurements, tabulate the data, and then you try to find a mathematical model that fits your numerical data.
Fortunately, there is a simple test to check if a numerically given function is exponential.

Example 5.3.1.

Consider the exponential function \(y = f(x) = 3 \cdot 2^x\text{.}\) For a few equally spaced values of \(x\text{,}\) here are the values from the function:
\(x\) \(0\) \(1\) \(2\) \(3\) \(4\) \(5\)
\(y\) \(3\) \(6\) \(12\) \(24\) \(48\) \(96\)
The values of \(x\) are equally spaced; that is, the difference between two consecutive values of \(x\) is constantly equal to \(\Delta x = 1\text{:}\)
\begin{equation*} 1 - 0 = 1, \quad 2 - 1 = 1, \quad 3 - 2 = 1, \quad 4 - 3 = 1, \quad 5 - 4 = 1. \end{equation*}
The differences of the corresponding \(y\)-values are, of course, not equal. If the differences of consecutive values of \(y\) were equal, the table would represent a linear function. Instead, the ratios of consecutive values of \(y\) are equal:
\begin{equation*} \dfrac{6}{3} = 2, \quad \dfrac{12}{6} = 2, \quad \dfrac{24}{12} = 2, \quad \dfrac{48}{24} = 2, \quad \dfrac{96}{48} = 2. \end{equation*}
Of course, \(f(x) = 3 \cdot 2^x\) is an exponential functionβ€”it matches the form \(y = A\cdot b^x\text{.}\) This example shows that the property of constant ratios between consecutive outputs indicates that a function is exponential.
For every exponential function and equally spaced values of \(x\text{,}\) the ratios of the consecutive values of \(y\) are equal. The previous example illustrates this point, but it can also be seen from algebraic properties of exponential expressions. Take any exponential function \(y = f(x) = A \cdot b^x\text{.}\) Consider a few equally spaced values for \(x\text{,}\) denoted by \(x_1, x_2, x_3, x_4\text{,}\) meaning that
\begin{equation*} x_2 - x_1 = x_3 - x_2 = x_4 - x_3 = \Delta x. \end{equation*}
where \(\Delta x\) is some constant number. The corresponding \(y\)-values are \(f(x_1) = A \cdot b^{x_1}\text{,}\) \(f(x_2) = A \cdot b^{x_2}\text{,}\) and so on. The consecutive ratios between \(y\)-values are:
\begin{align*} \dfrac{A \cdot b^{x_2}}{A \cdot b^{x_1}} \amp = \dfrac{b^{x_2}}{b^{x_1}} = b^{x_2 - x_1} = b^{\Delta x}\\ \dfrac{A \cdot b^{x_3}}{A \cdot b^{x_2}} \amp = \dfrac{b^{x_3}}{b^{x_2}} = b^{x_3 - x_2} = b^{\Delta x}\\ \dfrac{A \cdot b^{x_4}}{A \cdot b^{x_3}} \amp= \dfrac{b^{x_4}}{b^{x_3}} = b^{x_4 - x_3} = b^{\Delta x} \end{align*}
All the ratios are equal to \(b^{\Delta x}\text{,}\) which is a constant number.
Note: If the values of \(x\) are spaced by \(1\) as in ExampleΒ 5.3.1β€”that is, \(\Delta x = 1\)β€”the ratios give \(b\text{,}\) the base of the exponential function. If the \(x\) values are spaced by \(2\) or any other distance, so that \(\Delta x \neq 1\text{,}\) the ratios between consecutive \(y\)-values are not equal to \(b\text{!}\)

Example 5.3.2.

Which of the functions given below are exponential? Give a formula for each function.
\(x\) \(0\) \(1\) \(2\) \(3\) \(4\) \(5\)
\(y\) \(0.5\) \(2\) \(8\) \(32\) \(128\) \(512\)
\(x\) \(0\) \(1\) \(2\) \(3\) \(4\) \(5\)
\(y\) \(0.5\) \(4.5\) \(8.5\) \(12.5\) \(16.5\) \(20.5\)
\(x\) \(1\) \(2\) \(3\) \(4\) \(5\)
\(y\) \(2\) \(1\) \(0.5\) \(0.25\) \(0.125\)
Solution.
(a) The values of \(x\) are equally spaced, namely by \(\Delta x = 1\text{.}\) We have to check if all ratios between consecutive values of \(y\) are the same:
\begin{equation*} \dfrac{2}{0.5} = 4, \quad \dfrac{8}{2} = 4, \quad \dfrac{32}{8} = 4, \quad \dfrac{128}{32} = 4, \quad \dfrac{512}{128} = 4. \end{equation*}
All ratios are the same. Since \(\Delta x = 1\text{,}\) the ratios give us the growth factor \(b = 4\text{.}\) With each increase in \(x\) by \(1\text{,}\) the current value of the function is multiplied by the constant factor of \(4\text{.}\) The initial value is the \(y\) value at \(0\text{,}\) so \(A = 0.5\text{.}\) The table corresponds to the exponential function \(y = 0.5 \cdot 4^x\text{.}\)
(b) The ratios between consecutive outputs are
\begin{equation*} \dfrac{4.5}{0.5} = 9, \quad \dfrac{8.5}{4.5} = 1.888, \qquad \dots \end{equation*}
which are clearly not equal, so the function is not exponential. Observe that the values of the function start at \(y=0.5\) when \(x = 0\) and then increase by \(4\) for every unit increase in \(x\text{.}\) This is precisely the definition of a linear function, so table (b) corresponds to \(y = 0.5 + 4x\text{.}\)
(c) The ratios between consecutive outputs are
\begin{equation*} \dfrac{1}{2} = 0.5, \quad \dfrac{0.5}{1} = 0.5, \quad \dfrac{0.25}{0.5} = 0.5, \quad \dfrac{0.125}{0.25} = 0.5 \end{equation*}
which are equal, and the inputs are evenly spaced by \(\Delta x = 1\text{,}\) so the function is exponential. The growth factor is \(b = 0.5\) and so \(y = A \cdot 0.5^x\text{.}\) The initial value \(A\) is unknown. Because the value of the function at \(0\) is known, set up an equation for \(A\) using, for example, the point \((1, 2)\) from the table. This gives the equation \(A \cdot 0.5^1 = 2\text{,}\) and so \(A = 4\text{.}\) The function represented by the table (c) is \(y = 4 \cdot 0.5^x\text{.}\)

Example 5.3.3.

Check if \(f(t)\) is an exponential function. If so, find the formula for \(f(t)\text{.}\)
\(t\) \(0\) \(2\) \(4\) \(6\) \(8\) \(10\)
\(f(t)\) \(0.3\) \(0.9\) \(2.7\) \(8.1\) \(24.3\) \(72.9\)
Solution.
The values of \(t\) are equally spaced by \(2\text{;}\) that is, \(\Delta t = 2\text{.}\) The ratios between consecutive output values are:
\begin{equation*} \dfrac{0.9}{0.3} = 3, \quad \dfrac{2.7}{0.9} = 3, \quad \dfrac{8.1}{2.7} = 3, \quad \dfrac{24.3}{8.1} = 3, \quad \dfrac{72.9}{24.3} = 3. \end{equation*}
These ratios are all equal, so \(f(t) = A \cdot b^t\) is an exponential function.
The initial value is provided from the table as \(A = f(0) = 0.3\text{.}\) This means \(f(t) = 0.3 \cdot b^t\text{.}\) Now we need to find the growth factor \(b\text{.}\) Is \(b\) equal to \(3\text{?}\) No! At each step \(t\) changes by \(2\) units, so the current value of \(f(t)\) is multiplied by \(b\) twice. To find \(b\text{,}\) use any pair of values from the table to set up an equation. For example, using \(f(2) = 0.9\text{,}\) then:
\begin{align*} f(2) \amp = 0.9\\ 0.3 \cdot b^2 \amp = 0.9\\ b^2 \amp = 0.9 / 0.3\\ b^2 \amp = 3\\ b \amp = \sqrt{3} \end{align*}
Remember that \(b\) has to be positive so \(b = \sqrt{3}\) rather than \(b = \pm \sqrt{3}\text{.}\) The function is \(f(t) = 0.3 \cdot (\sqrt{3})^t\text{.}\) Using \(\sqrt{3} \approx 1.732\text{,}\) then \(f(t) = 0.3 (1.732)^t\text{.}\) We can also rewrite \(f(t)\) as \(f(t) = 0.3 (3)^{\frac{t}{2}}\text{,}\) which shows that the values of the function triple every two units of \(t\text{.}\)

Subsection Modeling: The Population of Mozambique

Example 5.3.4.

The population of Mozambique
 1 
https://www.worldometers.info/world-population/mozambique-population/, accessed: 3/10/2020
between 2009 and 2014 is given below:
Year \(2009\) \(2010\) \(2011\) \(2012\) \(2013\) \(2014\)
Pop. in millions \(22.89\) \(23.53\) \(24.19\) \(24.86\) \(25.56\) \(26.28\)
Find a function that models Mozambique’s population growth in the time period 2009-2014.
Solution.
To simplify the task, denote by \(t\) the number of years since 2009 and by \(P(t)\) the population at time \(t\) in millions. The table then becomes:
\(t\) \(0\) \(1\) \(2\) \(3\) \(4\) \(5\)
\(P(t)\) \(22.89\) \(23.53\) \(24.19\) \(24.86\) \(25.56\) \(26.28\)
Populations often increase exponentially so it is reasonable to check if \(P(t)\) is exponential. To do so, check the ratios between consecutive outputs in the table:
\begin{equation*} \dfrac{23.53}{22.89} \approx 1.028, \quad \dfrac{24.19}{23.53} \approx 1.028, \quad \dfrac{24.86}{24.19} \approx 1.028, \quad \dfrac{25.56}{24.86} \approx 1.028, \quad \dfrac{26.28}{25.56} \approx 1.028. \end{equation*}
When rounded to three decimal places, the ratios are equal so \(P(t)\) can be reasonably approximated by an exponential function. Because the inputs are evenly spaced by \(1\text{,}\) the corresponding growth factor would be \(b = 1.028\text{.}\) The initial value is immediately given in the table as \(A = P(0) = 22.89\text{.}\) Therefore, the function that approximately models the population growth in Mozambique is
\begin{equation*} P(t) = 22.89 (1.028)^t. \end{equation*}
As the growth factor is \(1.028\text{,}\) the annual percent growth rate is \(r=2.8\%\text{.}\) The model is valid in the time period 2009-2014. If you examine the data given at worldometers.info website, the formula is valid for a longer time interval. The annual percent growth rate remains at approximately \(2.8\%\) between 2005 and 2020.

Exercises Exercises

Exponential Functions.

For each of the following data tables, check if the data can possibly correspond to an exponential function or not. If yes, give a formula for the function in the form \(A\cdot b^t\text{.}\) Identify the initial value and the growth factor.

1.

\(t\) \(0\) \(1\) \(2\) \(3\) \(4\) \(5\)
\(f(t)\) \(5000\) \(4000\) \(3200\) \(2560\) \(2048\) \(1638.4\)
Solution.
possible exponential function; \(f(t)=5000(0.8)^t\text{;}\) initial value: \(5000\text{;}\) growth factor: \(0.8\)

2.

\(t\) \(0\) \(1\) \(2\) \(3\) \(4\) \(5\)
\(g(t)\) \(32\) \(48\) \(72\) \(108\) \(162\) \(243\)
Solution.
possible exponential function; \(g(t)=32(1.5)^t\text{;}\) initial value: \(32\text{;}\) growth factor: \(1.5\)

3.

\(t\) \(0\) \(1\) \(2\) \(3\) \(4\) \(5\)
\(k(t)\) \(64\) \(32\) \(16\) \(8\) \(0\) \(2\)
Solution.
not an exponential function

4.

\(t\) \(1\) \(2\) \(3\) \(4\) \(5\)
\(h(t)\) \(1280\) \(1600\) \(2000\) \(2500\) \(3125\)
Solution.
possible exponential function; \(h(t)=1024(1.25)^t\text{;}\) initial value: \(1024\text{;}\) growth factor: \(1.25\)

5.

\(t\) \(0\) \(1\) \(2\) \(3\) \(4\) \(5\)
\(n(t)\) \(100\) \(200\) \(300\) \(400\) \(500\) \(600\)
Solution.
not an exponential function

Exponential Functions.

For each of the following data tables, check if the data can possibly correspond to an exponential function or not. If yes, give a formula for the function in the form \(A\cdot b^t\text{.}\) Give both the exact and the approximate value for the growth factor. Round off to at least three decimal places.

6.

\(t\) \(0\) \(2\) \(4\) \(6\) \(8\)
\(z(t)\) \(250\) \(500\) \(1000\) \(2000\) \(4000\)
Solution.
possible exponential function; \(z(t)=250(\sqrt{2})^t\approx250(1.414)^t\)

7.

\(t\) \(0\) \(3\) \(6\) \(9\) \(12\) \(15\)
\(m(t)\) \(64\) \(32\) \(16\) \(8\) \(4\) \(2\)
Solution.
possible exponential function; \(m(t)=64\left(\sqrt[3]{\frac{1}{2}}\right)^t\approx64(0.794)^t\)

Finding a Function.

For each of the following, find a function \(y=f(x)\) that corresponds to the data given in each table. Fill in the missing entries.

8.

\(x\) \(0\) \(1\) \(2\) \(3\) \(4\)
\(y\) \(500\) \(900\) \(1620\) \(2916\) \(?\)
Solution.
\(y=500 (1.8)^x\text{;}\) \(y(4)=5248.8\)

9.

\(x\) \(0\) \(1\) \(2\) \(3\) \(4\)
\(y\) \(300\) \(220\) \(140\) \(60\) \(?\)
Solution.
\(y=-80x+300\text{;}\) \(y(4)=-20\)

10.

Barometric pressure
 2 
β€”the pressure of the airβ€”depends on the altitude above sea level. Let \(H\) be altitude above sea level measured in kilometers. Let \(P(H)\) be barometric pressure measured in mmHgβ€”millimeters of mercury. Here are readings of barometric pressure at different altitudes:
\(H\) (km) \(0\) \(1\) \(2\) \(3\) \(4\)
\(P(H)\) (mmHg) \(760\) \(674.12\) \(597.94\) \(530.37\) \(470.44\)

(c)

Find a formula for \(P(H)\) in the form \(P(H)=P_0\cdot b^H\text{.}\) When calculating \(b\text{,}\) round off to three decimal places.
Solution.
\(P(H)=760(0.887)^H\)

(d)

The summit of Mount Everest is at \(8848\) meters above sea level. What is the barometric pressure at the top of Mount Everest?
Solution.
approximately \(263.05\) mmHg
You have attempted of activities on this page.