Calculus3Analyzing Functions1 of 6
Analyzing Functions · 1

Extreme Values on a Closed Interval

1

Absolute maximum and minimum values

Let's look at f(x)=x2f(x) = x^2 on the closed interval [−1,2]\iv{[-1, 2]}. What's the largest value of f(x)f(x), and what's the smallest?

The largest value ff takes on an interval is its absolute maximum value. The smallest value is its absolute minimum value. Together, we call them the extreme values of ff on the interval.

11
22
33
44
xx
yy
−1\htmlClass{lc-orange}{-1}
2\htmlClass{lc-orange}{2}
f(−1)=1f(-1)=1
f(0)=0\htmlClass{lc-blue}{f(0)=0}
f(2)=4\htmlClass{lc-pink}{f(2)=4}
absolute maximum value=f(2)=4absolute minimum value=f(0)=0\displaystyle \begin{aligned} &\mx{\text{absolute maximum value}} \\ &\qquad \mx{= f(2) = 4} \\[8pt] &\mn{\text{absolute minimum value}} \\ &\qquad \mn{= f(0) = 0} \end{aligned}

Notice that x=−1\iv{x = -1} is an endpoint, but f(−1)=1f(-1) = 1 is less than f(2)=4\mx{f(2) = 4}. So an endpoint doesn't have to give the largest value.

The value itself is the y-coordinate, and where it happens is the x-coordinate. So we say the absolute maximum value is 4\mx{4}, at x=2\mx{x = 2}, and the absolute minimum value is 0\mn{0}, at x=0\mn{x = 0}.

2

The Extreme Value Theorem

Does every function on an interval have an absolute maximum value and an absolute minimum value? Not always. But if we keep the interval closed and keep the function continuous, it has to have both.

The Extreme Value Theorem
f continuous on [a,b]⟹    f has an absolute maximum valueand an absolute minimum value on [a,b]\displaystyle \begin{aligned} &f \text{ continuous on } \iv{[a, b]} \\ \Longrightarrow\;\; &f \text{ has an } \mx{\text{absolute maximum value}} \\ &\text{and an } \mn{\text{absolute minimum value}} \text{ on } \iv{[a, b]} \end{aligned}
f continuous on [a,b]⟹  f has anabsolute maximum valueand an absolute minimum valueon [a,b]\displaystyle \begin{gathered} f \text{ continuous on } \iv{[a, b]} \\ \Longrightarrow\; f \text{ has an} \\ \mx{\text{absolute maximum value}} \\ \text{and an } \mn{\text{absolute minimum value}} \\ \text{on } \iv{[a, b]} \end{gathered}

Let's see why we need both conditions, and why a sharp corner doesn't matter.

Finding when a function is guaranteed an absolute maximum and minimum
11
22
33
44
xx
yy
−1\htmlClass{lc-orange}{-1}
2\htmlClass{lc-orange}{2}
f(x)=x2, [−1,2]f(x)=x^2,\ \htmlClass{lc-orange}{[-1,2]}
Text solution
Leave out an endpointOn [−1,2)\iv{[-1, 2)}, x2x^2 is always less than 44 and never equals it. Whatever xx we pick, an xx closer to 22 gives a larger value, so there's no largest value.
f(1.9)=3.61,f(1.95)=3.8025\displaystyle \tp{f(1.9) = 3.61}, \qquad \mx{f(1.95) = 3.8025}
Break the functionOn [−1,1]\iv{[-1, 1]}, 1x2\frac{1}{x^2} isn't defined at 00, so it isn't continuous there, and near 00 its values keep growing.
x→0 ⁣:1x2→∞\displaystyle x \to 0\colon \quad \frac{1}{x^2} \to \infty
A corner is still continuous∣x−1∣|x - 1| has a corner at x=1x = 1, but it's continuous on [0,3]\iv{[0, 3]}, so it has both extreme values.
absolute maximum value 2\displaystyle \mx{\text{absolute maximum value } 2}
absolute minimum value 0\displaystyle \mn{\text{absolute minimum value } 0}

With the right endpoint left out, the point (2,4)(2, 4) is missing from the graph, and the values climb toward 44 without ever reaching it.

11
22
33
44
xx
yy
−1\htmlClass{lc-orange}{-1}
2\htmlClass{lc-orange}{2}
x2  on [−1,2) ⁣:f(x)<4\displaystyle x^2 \ \text{ on } \iv{[-1, 2)}\colon \quad f(x) < 4
f(1.9)=3.61f(1.95)=3.8025\displaystyle \begin{aligned} \tp{f(1.9)} &= \tp{3.61} \\ \mx{f(1.95)} &= \mx{3.8025} \end{aligned}

Can you find a largest value when the right endpoint is excluded?

A closed interval doesn't help if the function has a break. Near 00, the graph of 1x2\frac{1}{x^2} goes up without end.

xx
11
55
−1\htmlClass{lc-orange}{-1}
1\htmlClass{lc-orange}{1}
f(−1)=1f(-1)=1
f(1)=1f(1)=1
f(x)\displaystyle f(x) =1x2  on [−1,1]\displaystyle = \frac{1}{x^2} \ \text{ on } \iv{[-1, 1]}
f(0) undefined\displaystyle f(0) \text{ undefined}
x→0 ⁣:1x2→∞\displaystyle \begin{gathered} x \to 0\colon \quad \frac{1}{x^2} \to \infty \end{gathered}

A graph doesn't need to be smooth, though. The corner of ∣x−1∣|x - 1| at x=1x = 1 has no derivative, but the graph has no break there.

11
22
xx
yy
0\htmlClass{lc-orange}{0}
3\htmlClass{lc-orange}{3}
f(1)=0\htmlClass{lc-blue}{f(1)=0}
f(3)=2\htmlClass{lc-pink}{f(3)=2}
f(x)\displaystyle f(x) =∣x−1∣  on [0,3]\displaystyle = |x - 1| \ \text{ on } \iv{[0, 3]}
f′(1) doesn’t exist\displaystyle \slp{f'(1) \text{ doesn't exist}}
absolute maximum value 2,\displaystyle \mx{\text{absolute maximum value } 2,}
at x=3\displaystyle \mx{\text{at } x = 3}
absolute minimum value 0,\displaystyle \mn{\text{absolute minimum value } 0,}
at x=1\displaystyle \mn{\text{at } x = 1}

The Extreme Value Theorem guarantees that an absolute maximum and an absolute minimum exist, but it doesn't tell us what those values are or which x-values give them.

3

Where the highest and lowest points can be

So how do we actually find them? Let's find the largest and smallest values of f(x)=2x3−3x2−12xf(x) = 2x^3 - 3x^2 - 12x on the interval [−2,4]\iv{[-2, 4]}.

f(x)=2x3−3x2−12x  on [−2,4]largest value=  ?smallest value=  ?\displaystyle \begin{gathered} f(x) = 2x^3 - 3x^2 - 12x \ \text{ on } \iv{[-2, 4]} \\ \mx{\text{largest value} = \;?} \qquad \mn{\text{smallest value} = \;?} \end{gathered}

We could try plugging in integers.

x−2−101234f(x)−470−13−20−932\displaystyle \begin{array}{c|ccccccc} x & -2 & -1 & 0 & 1 & 2 & 3 & 4 \\ \hline f(x) & -4 & 7 & 0 & -13 & -20 & -9 & 32 \end{array}
xf(x)−2−4−17001−132−203−9432\displaystyle \begin{array}{c|c} x & f(x) \\ \hline -2 & -4 \\ -1 & 7 \\ 0 & 0 \\ 1 & -13 \\ 2 & -20 \\ 3 & -9 \\ 4 & 32 \end{array}

But these are only seven of infinitely many x-values. The function could be greater than 3232 somewhere between 33 and 44, and a table can't tell us. So instead, let's ask where the highest and lowest points can't be.

Find where a curve on [a,b][a, b] can have its highest and lowest points
a\htmlClass{lc-orange}{a}
b\htmlClass{lc-orange}{b}
largest value=?smallest value=?\begin{aligned}\htmlClass{lc-pink}{\text{largest value}} &\htmlClass{lc-pink}{{}={}?}\\ \htmlClass{lc-blue}{\text{smallest value}} &\htmlClass{lc-blue}{{}={}?}\end{aligned}
Text solution
Where f′(x)≠0\slp{f'(x) \ne 0}, rule the point outZoomed in, the curve looks like its tangent line. If x\tp{x} is inside the interval and the slope there isn't 00, the curve is higher on one side of x\tp{x} and lower on the other, so x\tp{x} can't be the highest or the lowest point.
f′(x)>0 ⁣: higher to the right\displaystyle \slp{f'(x) > 0}\colon \ \text{higher to the right}
f′(x)<0 ⁣: higher to the left\displaystyle \slp{f'(x) < 0}\colon \ \text{higher to the left}
What's left insideThe argument needs a tangent line with a nonzero slope, so two kinds of points stay on the list: a horizontal tangent, and a sharp corner, where there's no tangent line at all.
f′(x)=0\displaystyle \slp{f'(x) = 0}
f′(x) doesn’t exist\displaystyle \slp{f'(x) \text{ doesn't exist}}
The endpoints stayIf f′(b)>0\slp{f'(b) > 0}, the higher points just to the right of b\iv{b} are outside the interval, so b\iv{b} can still be the highest point. The same goes for a\iv{a}, whatever the slope.

Here's the picture to keep in mind. Where f′(x)>0\slp{f'(x) > 0}, the curve is higher just to the right of x\tp{x} and lower just to the left, so x\tp{x} can't be the highest or the lowest point.

a\htmlClass{lc-orange}{a}
b\htmlClass{lc-orange}{b}
x\htmlClass{lc-violet}{x}
f′(x)>0\htmlClass{lc-green}{f'(x)>0}

The points inside the interval that stay on the list have a name. A critical point of ff is an xx-value where f(x)f(x) is defined and f′(x)=0\slp{f'(x) = 0} or f′(x)\slp{f'(x)} doesn't exist.

f(x) defined,f′(x)=0  or  f′(x) doesn’t exist\displaystyle f(x) \text{ defined}, \quad \slp{f'(x) = 0} \ \text{ or } \ \slp{f'(x) \text{ doesn't exist}}
f(x) defined,f′(x)=0  or  f′(x) doesn’t exist\displaystyle \begin{gathered} f(x) \text{ defined}, \\ \slp{f'(x) = 0} \ \text{ or } \ \slp{f'(x) \text{ doesn't exist}} \end{gathered}

So for a continuous function on [a,b]\iv{[a, b]}, the absolute maximum and minimum can only occur at three kinds of points.

Candidates for highest and lowest points
a\htmlClass{lc-orange}{a}
b\htmlClass{lc-orange}{b}
f′(x)=0\htmlClass{lc-green}{f'(x)=0}
f′(x) doesn’t exist\htmlClass{lc-green}{f'(x)\text{ doesn't exist}}
Endpoints
x=a,x=b\displaystyle \iv{x = a}, \qquad \iv{x = b}
Horizontal tangent
f′(x)=0\displaystyle \slp{f'(x) = 0}
Sharp point
f′(x) doesn’t exist,f(x) defined\displaystyle \slp{f'(x) \text{ doesn't exist}}, \quad f(x) \text{ defined}

A critical point is a candidate, not an answer. The top of a hill, where f′(x)=0\slp{f'(x) = 0}, can still be lower than an endpoint. So we find the value of ff at every candidate and compare them. We don't need to work out what kind of critical point each candidate is.

4

Testing the candidates on a closed interval

To find the absolute maximum and minimum values of a continuous function on a closed interval [a,b]\iv{[a, b]}, we only need to test the endpoints and any critical points inside the interval.

The closed interval method
Condition
f continuous on [a,b]\displaystyle f \text{ continuous on } \iv{[a, b]}
Find critical points
critical points of f in [a,b]\displaystyle \slp{\text{critical points}} \text{ of } f \text{ in } \iv{[a, b]}
Evaluate
f(a),f(b)\displaystyle f(\iv{a}), \qquad f(\iv{b})
f at each critical point\displaystyle f \text{ at each } \slp{\text{critical point}}
Compare
largest=absolute maximum value\displaystyle \mx{\text{largest} = \text{absolute maximum value}}
smallest=absolute minimum value\displaystyle \mn{\text{smallest} = \text{absolute minimum value}}

The condition matters. For f(x)=1x2f(x) = \frac{1}{x^2} on [−1,1]\iv{[-1, 1]}, there are no critical points, so the method would test only the endpoints and call 11 the maximum. But near 00 the values go to infinity. ff isn't continuous at 00, so the theorem doesn't apply.

f(x)\displaystyle f(x) =1x2  on [−1,1]\displaystyle = \frac{1}{x^2} \ \text{ on } \iv{[-1, 1]}
f′(x)=−2x3≠0\displaystyle \slp{f'(x) = -\frac{2}{x^3} \neq 0}
f(−1)\displaystyle f(-1) =f(1)\displaystyle = f(1) =1,f(0.1)=100\displaystyle = 1, \qquad \mx{f(0.1) = 100}

Let's use the method on the cubic from the last section, first on [−2,4]\iv{[-2, 4]} and then on a smaller interval.

Find the absolute maximum and minimum of f(x)=2x3−3x2−12xf(x) = 2x^3 - 3x^2 - 12x on [−2,4][-2, 4], and then on [0,3][0, 3]
f(x)\displaystyle f(x) =2x3−3x2−12x\displaystyle = 2x^3 - 3x^2 - 12x
largestvaluesmallestvalue[−2,4]??[0,3]??\displaystyle \def\arraystretch{1.3}\begin{array}{c|c|c} & \htmlClass{lc-pink}{\begin{matrix}\text{largest}\\[-3pt]\text{value}\end{matrix}} & \htmlClass{lc-blue}{\begin{matrix}\text{smallest}\\[-3pt]\text{value}\end{matrix}} \\ \hline \htmlClass{lc-orange}{[-2, 4]} & ? & ? \\ \htmlClass{lc-orange}{[0, 3]} & ? & ? \end{array}
Text solution
Find the critical pointsff is a polynomial, so it's continuous and its derivative exists everywhere. The only critical points are where f′(x)=0\slp{f'(x) = 0}.
f′(x)\displaystyle \slp{f'(x)} =6(x−2)(x+1)\displaystyle = \slp{6(x-2)(x+1)}
x=−1,x=2\displaystyle \slp{x = -1}, \quad \slp{x = 2}
Candidates on [−2,4]\iv{[-2, 4]}Both critical points are inside, so we test four values: the endpoints −2\iv{-2} and 4\iv{4}, and −1\slp{-1} and 2\slp{2}.
f(−2)\displaystyle f(-2) =−4,f(−1)\displaystyle = -4, \quad f(-1) =7,f(2)=−20,f(4)=32\displaystyle = 7, \quad \mn{f(2) = -20}, \quad \mx{f(4) = 32}
Candidates on [0,3]\iv{[0, 3]}The critical points don't change, but −1\spent{-1} isn't in [0,3]\iv{[0, 3]}, so we drop it and test the endpoints 0\iv{0} and 3\iv{3}, and 2\slp{2}.
f(0)=0,f(2)=−20,f(3)\displaystyle \mx{f(0) = 0}, \quad \mn{f(2) = -20}, \quad f(3) =−9\displaystyle = -9

On [−2,4]\iv{[-2, 4]}, the largest value is 32\mx{32}, at the endpoint x=4\mx{x = 4}. That's higher than the top of the hill at x=−1x = -1, where f(−1)=7f(-1) = 7. If we had left out the endpoints, we would have picked 77 as the maximum.

xf(x)−2−4endpoint−17critical point2−20critical point432endpoint\displaystyle \def\arraystretch{1.3}\begin{array}{c|c|l} x & f(x) & \\ \hline -2 & -4 & \text{endpoint} \\ -1 & 7 & \text{critical point} \\ \mn{2} & \mn{{-20}} & \text{critical point} \\ \mx{4} & \mx{32} & \text{endpoint} \end{array}
absolute maximum value 32,\displaystyle \mx{\text{absolute maximum value } 32,}
at x=4\displaystyle \mx{\text{at } x = 4}
absolute minimum value −20,\displaystyle \mn{\text{absolute minimum value } {-20},}
at x=2\displaystyle \mn{\text{at } x = 2}

On [0,3]\iv{[0, 3]}, the maximum moves to the left endpoint. If we had kept −1\spent{-1}, its value 77 would have looked like the maximum, but f(x)≤0f(x) \le 0 for every xx in [0,3]\iv{[0, 3]}.

xf(x)−1not in [0,3]00endpoint2−20critical point3−9endpoint\displaystyle \def\arraystretch{1.3}\begin{array}{c|c|l} x & f(x) & \\ \hline \spent{-1} & & \spent{\text{not in } [0, 3]} \\ \mx{0} & \mx{0} & \text{endpoint} \\ \mn{2} & \mn{{-20}} & \text{critical point} \\ 3 & -9 & \text{endpoint} \end{array}
absolute maximum value 0,\displaystyle \mx{\text{absolute maximum value } 0,}
at x=0\displaystyle \mx{\text{at } x = 0}
absolute minimum value −20,\displaystyle \mn{\text{absolute minimum value } {-20},}
at x=2\displaystyle \mn{\text{at } x = 2}

Can an endpoint give the absolute maximum or minimum instead of a critical point?

5

Where the derivative doesn't exist

For the cubic, every critical point came from f′(x)=0\slp{f'(x) = 0}. In the next example the derivative is never 00, and that makes a critical point easy to miss.

Find the absolute maximum and minimum of f(x)=x23f(x) = x^{\frac23} on [−1,8][-1, 8]
f(x)\displaystyle f(x) =x23,[−1,8]\displaystyle = x^{\frac23}, \quad \htmlClass{lc-orange}{[-1, 8]}
largest value=?smallest value=?\displaystyle \begin{aligned}\htmlClass{lc-pink}{\text{largest value}} &\htmlClass{lc-pink}{{}={}?}\\ \htmlClass{lc-blue}{\text{smallest value}} &\htmlClass{lc-blue}{{}={}?}\end{aligned}
Text solution
The derivative is never 00The top of the fraction is 2\slp{2}, so f′(x)=0\slp{f'(x) = 0} has no solution.
f′(x)\displaystyle \slp{f'(x)} =23x−13\displaystyle = \slp{\frac23 x^{-\frac13}} =23x3\displaystyle = \slp{\frac{2}{3\sqrt[3]{x}}}
Check where the bottom is 00At x=0\slp{x = 0} we'd divide by 00, so f′(0)\slp{f'(0)} doesn't exist. But ff is defined there, so x=0\slp{x = 0} is a critical point.
3x3=0  →  x=0\displaystyle \slp{3\sqrt[3]{x} = 0} \;\to\; \slp{x = 0}
f(0)=0\displaystyle \mn{f(0) = 0}
CompareThe endpoints alone would make 11 look like the minimum, but f(0)=0\mn{f(0) = 0} is smaller.
f(−1)\displaystyle f(-1) =1,f(0)=0,f(8)=4\displaystyle = 1, \quad \mn{f(0) = 0}, \quad \mx{f(8) = 4}

On the graph, the critical point is the sharp point at (0,0)\mn{(0, 0)}, below both endpoints.

44
xx
yy
−1\htmlClass{lc-orange}{-1}
8\htmlClass{lc-orange}{8}
f(−1)=1f(-1)=1
f(0)=0\htmlClass{lc-blue}{f(0)=0}
f(8)=4\htmlClass{lc-pink}{f(8)=4}

With the endpoints and the critical point compared, the largest value is 4\mx{4}, and the smallest is 0\mn{0}.

xf(x)−11endpoint00critical point84endpoint\displaystyle \def\arraystretch{1.3}\begin{array}{c|c|l} x & f(x) & \\ \hline -1 & 1 & \text{endpoint} \\ \mn{0} & \mn{0} & \text{critical point} \\ \mx{8} & \mx{4} & \text{endpoint} \end{array}
absolute maximum value 4,\displaystyle \mx{\text{absolute maximum value } 4,}
at x=8\displaystyle \mx{\text{at } x = 8}
absolute minimum value 0,\displaystyle \mn{\text{absolute minimum value } 0,}
at x=0\displaystyle \mn{\text{at } x = 0}

Absolute values have the same kind of critical point. For ∣x2−4∣|x^2 - 4|, the graph has corners where x2−4=0\slp{x^2 - 4 = 0}, so the derivative doesn't exist at x=−2\slp{x = -2} and x=2\slp{x = 2}.

xx
yy
x=−2\htmlClass{lc-green}{x=-2}
x=2\htmlClass{lc-green}{x=2}
4−x24-x^2
y=∣x2−4∣y=|x^2-4|
x2−4=0→  x=−2,x=2\displaystyle \begin{gathered} \slp{x^2 - 4 = 0} \\ \to\; \slp{x = -2}, \quad \slp{x = 2} \end{gathered}
6

Finding every critical point

In this example, f′(x)=0\slp{f'(x) = 0} has two solutions in the interval, and either one could give an extreme value, so we have to find both.

Find the absolute maximum and minimum of f(x)=x+2cos⁡xf(x) = x + 2\cos x on [0,π][0, \pi]
f(x)\displaystyle f(x) =x+2cos⁡x,[0,π]\displaystyle = x + 2\cos x, \quad \htmlClass{lc-orange}{[0, \pi]}
largest value=?smallest value=?\displaystyle \begin{aligned}\htmlClass{lc-pink}{\text{largest value}} &\htmlClass{lc-pink}{{}={}?}\\ \htmlClass{lc-blue}{\text{smallest value}} &\htmlClass{lc-blue}{{}={}?}\end{aligned}
Text solution
Set the derivative to 00The derivative exists everywhere, so the critical points come from f′(x)=0\slp{f'(x) = 0}.
f′(x)=1−2sin⁡x\displaystyle \slp{f'(x) = 1 - 2\sin x}
sin⁡x=12\displaystyle \slp{\sin x = \frac12}
Find both anglesSine is the height on the unit circle, and [0,π]\iv{[0, \pi]} is its top half. The height is 12\slp{\frac12} at two angles there.
x=π6,x=5π6\displaystyle \slp{x = \frac{\pi}{6}}, \qquad \slp{x = \frac{5\pi}{6}}
Compare with decimalsEvaluate ff at both critical points and both endpoints, keeping in mind that cos⁡5π6\cos\frac{5\pi}{6} is negative. Then compare with π≈3.14\pi \approx 3.14 and 3≈1.73\sqrt3 \approx 1.73.
f(π6)=π6+3≈2.26\displaystyle \mx{f\left(\frac{\pi}{6}\right) = \frac{\pi}{6} + \sqrt3 \approx 2.26}
f(5π6)=5π6−3≈0.89\displaystyle \mn{f\left(\frac{5\pi}{6}\right) = \frac{5\pi}{6} - \sqrt3 \approx 0.89}

On the unit circle, both angles have height 12\slp{\frac12} on the top half, one in the first quadrant and one in the second.

π6\htmlClass{lc-green}{\dfrac{\pi}{6}}
5π6\htmlClass{lc-green}{\dfrac{5\pi}{6}}
12\htmlClass{lc-green}{\tfrac{1}{2}}
0\htmlClass{lc-orange}{0}
π\htmlClass{lc-orange}{\pi}
sin⁡x=12,0≤x≤π\displaystyle \slp{\sin x = \frac12}, \quad \iv{0 \le x \le \pi}
x=π6,x=5π6\displaystyle \slp{x = \frac{\pi}{6}}, \qquad \slp{x = \frac{5\pi}{6}}

The decimals only decide the order. The answers stay exact.

xf(x)≈022π6π6+32.265π65π6−30.89ππ−21.14\displaystyle \def\arraystretch{1.3}\begin{array}{c|c|c} x & f(x) & \approx \\ \hline 0 & 2 & 2 \\[8pt] \mx{\dfrac{\pi}{6}} & \mx{\dfrac{\pi}{6} + \sqrt3} & \mx{2.26} \\[12pt] \mn{\dfrac{5\pi}{6}} & \mn{\dfrac{5\pi}{6} - \sqrt3} & \mn{0.89} \\[8pt] \pi & \pi - 2 & 1.14 \end{array}
absolute maximum value π6+3,\displaystyle \mx{\text{absolute maximum value } \frac{\pi}{6} + \sqrt3,}
at x=π6\displaystyle \mx{\text{at } x = \frac{\pi}{6}}
absolute minimum value 5π6−3,\displaystyle \mn{\text{absolute minimum value } \frac{5\pi}{6} - \sqrt3,}
at x=5π6\displaystyle \mn{\text{at } x = \frac{5\pi}{6}}

This time both extreme values are at critical points inside the interval. If we had missed 5π6\slp{\frac{5\pi}{6}}, the smallest value on our list would have been f(π)=π−2≈1.14f(\pi) = \pi - 2 \approx 1.14, and we'd have the wrong minimum.

Summary

Absolute maximum and minimum values
x2  on [−1,2] ⁣:absolute maximum value 4, at x=2x^2 \ \text{ on } \iv{[-1, 2]}\colon \quad \mx{\text{absolute maximum value } 4, \text{ at } x = 2}
x2  on [−1,2] ⁣:absolute maximum value 4,at x=2\begin{aligned} &x^2 \ \text{ on } \iv{[-1, 2]}\colon \\ &\mx{\text{absolute maximum value } 4,} \\ &\mx{\text{at } x = 2} \end{aligned}

The absolute maximum value is the largest value f(x)f(x) reaches on the interval, and the absolute minimum value is the smallest. The value itself is the y-coordinate, and where it happens is the x-coordinate.

The Extreme Value Theorem
f continuous on [a,b]  ⟹absolute maximum value and absolute minimum value on [a,b]\begin{aligned} &f \text{ continuous on } \iv{[a, b]} \;\Longrightarrow \\ &\mx{\text{absolute maximum value}} \text{ and } \mn{\text{absolute minimum value}} \text{ on } \iv{[a, b]} \end{aligned}
f continuous on [a,b]⟹absolute maximum valueand absolute minimum valueon [a,b]\begin{gathered} f \text{ continuous on } \iv{[a, b]} \\ \Longrightarrow \\ \mx{\text{absolute maximum value}} \\ \text{and } \mn{\text{absolute minimum value}} \\ \text{on } \iv{[a, b]} \end{gathered}

If ff is continuous on a closed interval [a,b]\iv{[a, b]}, it always has an absolute maximum value and an absolute minimum value on that interval. If the interval isn't closed, like x2x^2 on [−1,2)\iv{[-1, 2)}, or if ff isn't continuous, like 1x2\frac{1}{x^2} on [−1,1]\iv{[-1, 1]}, an extreme value might not exist. A sharp corner doesn't make the function discontinuous. ∣x−1∣|x - 1| on [0,3]\iv{[0, 3]} has a corner at x=1x = 1, and its maximum is 2\mx{2} and its minimum is 0\mn{0}.

Where extreme values can happen
a\htmlClass{lc-orange}{a}
b\htmlClass{lc-orange}{b}
f′(x)=0\htmlClass{lc-green}{f'(x)=0}
a\htmlClass{lc-orange}{a}
b\htmlClass{lc-orange}{b}
f′(x)\htmlClass{lc-green}{f'(x)}
doesn’t exist\htmlClass{lc-green}{\text{doesn't exist}}
a\htmlClass{lc-orange}{a}
b\htmlClass{lc-orange}{b}
x=b\htmlClass{lc-orange}{x=b}
x=a,x=b\iv{x = a}, \qquad \iv{x = b}
f′(x)=0\slp{f'(x) = 0}
f′(x) doesn’t exist\slp{f'(x) \text{ doesn't exist}}

Inside the interval, any point where f′(x)>0\slp{f'(x) > 0} or f′(x)<0\slp{f'(x) < 0} can't be the highest or lowest point, because the curve is higher on one side and lower on the other. That leaves only the endpoints a\iv{a} and b\iv{b}, and critical points inside [a,b]\iv{[a, b]}. At a critical point, f(x)f(x) is defined and either f′(x)=0\slp{f'(x) = 0} or f′(x)\slp{f'(x)} doesn't exist.

The closed-interval method
f continuous on [a,b]f \text{ continuous on } \iv{[a, b]}
critical points of f in [a,b]\slp{\text{critical points}} \text{ of } f \text{ in } \iv{[a, b]}
f(a),f(b)f(\iv{a}), \quad f(\iv{b})
f at each critical pointf \text{ at each } \slp{\text{critical point}}
largest=absolute maximum value\mx{\text{largest} = \text{absolute maximum value}}
smallest=absolute minimum value\mn{\text{smallest} = \text{absolute minimum value}}
largest=absolute maximum value\begin{aligned} &\mx{\text{largest}} \\ &\mx{= \text{absolute maximum value}} \end{aligned}
smallest=absolute minimum value\begin{aligned} &\mn{\text{smallest}} \\ &\mn{= \text{absolute minimum value}} \end{aligned}

To find the extreme values of a continuous function on [a,b]\iv{[a, b]}, find the critical points of ff and drop any that lie outside [a,b]\iv{[a, b]}. Then evaluate ff at the endpoints and at each remaining critical point. The largest value is the absolute maximum value, and the smallest is the absolute minimum value. Because we compare the values directly, we don't need to decide what kind of critical point each one is.

Critical points where the derivative doesn't exist
f′(x)=23x3 ⁣:x=0\slp{f'(x) = \dfrac{2}{3\sqrt[3]{x}}}\colon \quad \slp{x = 0}
∣x2−4∣ ⁣:x=−2,x=2|x^2 - 4|\colon \quad \slp{x = -2}, \quad \slp{x = 2}

Write f′(x)\slp{f'(x)} as a single fraction. Look where the denominator is 00 as well as where the numerator is 00. Where the denominator is 00, the fraction is undefined. If ff is defined there but its derivative isn't, that x-value is a critical point, as at x=0\slp{x = 0} for x23x^{\frac23}. For absolute value functions, check where the expression inside the absolute value is 00, like x=−2\slp{x = -2} and x=2\slp{x = 2} for ∣x2−4∣|x^2 - 4|.

Finding every critical point
sin⁡x=12,0≤x≤π ⁣:\slp{\sin x = \dfrac12}, \quad \iv{0 \le x \le \pi}\colon
x=π6,x=5π6\slp{x = \dfrac{\pi}{6}}, \quad \slp{x = \dfrac{5\pi}{6}}

Find every solution of f′(x)=0\slp{f'(x) = 0} inside the interval. Leaving off a critical point can make you miss the absolute maximum or minimum value. When values involve roots or π\pi, use decimals to compare them, and write the final answer with exact values.

7

Practice

Give exact values for answers, and use decimals only to compare.

1
Find the absolute maximum and minimum values of f(x)=x2−4x+1f(x) = x^2 - 4x + 1 on the interval [0,5][0, 5].

x=2x = 2 is one candidate. Which two endpoints do we also need to check on [0,5][0, 5]?

Set the derivative equal to 00 to find the critical points in [0,5]\iv{[0, 5]}.
f′(x)=2x−4=0  →  x=2\displaystyle \slp{f'(x) = 2x - 4 = 0} \;\to\; \slp{x = 2}
Evaluate ff at the critical point and at the endpoints 0\iv{0} and 5\iv{5}.
f(0)\displaystyle f(0) =1,f(2)=−3,f(5)=6\displaystyle = 1, \qquad \mn{f(2) = -3}, \qquad \mx{f(5) = 6}
The largest value is 6\mx{6} and the smallest is −3\mn{-3}.
absolute maximum value 6,\displaystyle \mx{\text{absolute maximum value } 6,}
at x=5\displaystyle \mx{\text{at } x = 5}
absolute minimum value −3,\displaystyle \mn{\text{absolute minimum value } {-3},}
at x=2\displaystyle \mn{\text{at } x = 2}
2
The graph of ff on [−2,4][-2, 4] is shown. Find the absolute maximum and minimum values of ff on [−2,4][-2, 4], and where they occur.
−3-3
−2-2
−1-1
11
22
33
44
−2-2
−1-1
11
22
33
44
xx
yy

At the bottom of the valley, f(2)=0f(2) = 0. Is that below f(−2)f(-2)?

Read the values at the endpoints and where the tangent line is horizontal.
endpoints: f(−2)=−3,f(4)\displaystyle \text{endpoints: } \mn{f(-2) = -3}, \quad f(4) =3\displaystyle = 3
f′=0:f(0)=4,f(2)\displaystyle \slp{f' = 0}: \quad \mx{f(0) = 4}, \quad f(2) =0\displaystyle = 0
The largest y-value is 4\mx{4} and the smallest is −3\mn{-3}.
absolute maximum value 4,\displaystyle \mx{\text{absolute maximum value } 4,}
at x=0\displaystyle \mx{\text{at } x = 0}
absolute minimum value −3,\displaystyle \mn{\text{absolute minimum value } {-3},}
at x=−2\displaystyle \mn{\text{at } x = -2}
3
For which function does the Extreme Value Theorem guarantee an absolute maximum value on the given interval?

Is the interval closed? Is the function continuous at every point in the interval? Does a sharp corner stop a function from being continuous?

If you got (A) Check x=2x = 2. The fraction 1x−2\frac{1}{x-2} is undefined there, and 22 is inside [0,3][0, 3].

If you got (B) Check the interval. The interval (−1,2)(-1, 2) doesn't include the endpoints −1-1 and 22.

If you got (D) Check x=π2x = \frac{\pi}{2}. The function tan⁡x\tan x is undefined there, and π2\frac{\pi}{2} is inside [0,π][0, \pi].

The Extreme Value Theorem requires the interval to be closed and the function to be continuous on that interval.
(A): 1x−2 undefined at x\displaystyle \text{(A)}: \ \frac{1}{x-2} \text{ undefined at } x =2,2 in [0,3]\displaystyle = 2, \quad 2 \text{ in } \iv{[0, 3]}
(B): (−1,2) without −1 and 2\displaystyle \text{(B)}: \ \iv{(-1, 2)} \text{ without } -1 \text{ and } 2
(D): tan⁡x undefined at x\displaystyle \text{(D)}: \ \tan x \text{ undefined at } x =π2,π2 in [0,π]\displaystyle = \frac{\pi}{2}, \quad \frac{\pi}{2} \text{ in } \iv{[0, \pi]}
The function ∣x−1∣|x - 1| is continuous on all of [0,3]\iv{[0, 3]}, so the theorem guarantees an absolute maximum value even though f′(1)\slp{f'(1)} doesn't exist.
(C): ∣x−1∣ continuous on [0,3],f′(1) doesn’t exist\displaystyle \text{(C)}: \ |x - 1| \text{ continuous on } \iv{[0, 3]}, \quad \slp{f'(1) \text{ doesn't exist}}
4
Find the absolute maximum and minimum values of f(x)=x23(x−5)f(x) = x^{\frac23}(x - 5) on the interval [−1,3][-1, 3].

Put the derivative over one denominator. Where is the denominator equal to 00?

To evaluate (−1)23(-1)^{\frac23}, take the cube root first, then square the result.

DifferentiateUse the product rule to find f′(x)\slp{f'(x)}, then write it over a common denominator.
f′(x)\displaystyle \slp{f'(x)} =23x−13(x−5)+x23\displaystyle = \slp{\frac23 x^{-\frac13}(x - 5) + x^{\frac23}} =2(x−5)3x3+3x3x3\displaystyle = \slp{\frac{2(x-5)}{3\sqrt[3]{x}} + \frac{3x}{3\sqrt[3]{x}}} =5x−103x3\displaystyle = \slp{\frac{5x - 10}{3\sqrt[3]{x}}} =5(x−2)3x3\displaystyle = \slp{\frac{5(x-2)}{3\sqrt[3]{x}}}
Find the critical pointsThe numerator is 00 at x=2\slp{x = 2}, and the denominator is 00 at x=0\slp{x = 0}, where f(0)f(0) is defined.
f′(x)=0: x=2\displaystyle \slp{f'(x) = 0}: \ \slp{x = 2}
f′(x) doesn’t exist: x=0,f(0)\displaystyle \slp{f'(x) \text{ doesn't exist}}: \ \slp{x = 0}, \quad f(0) =0\displaystyle = 0
Test the candidatesEvaluate ff at the critical points and endpoints, using decimals to compare the negative values.
f(−1)\displaystyle f(-1) =(−1)23(−6)\displaystyle = (-1)^{\frac23}(-6) =1⋅(−6)\displaystyle = 1 \cdot (-6) =−6\displaystyle = -6
f(0)=0,f(2)\displaystyle \mx{f(0) = 0}, \qquad f(2) =−343≈−4.76,f(3)\displaystyle = -3\sqrt[3]{4} \approx -4.76, \qquad f(3) =−293≈−4.16\displaystyle = -2\sqrt[3]{9} \approx -4.16
The largest value is 0\mx{0} and the smallest is −6\mn{-6}.
absolute maximum value 0,\displaystyle \mx{\text{absolute maximum value } 0,}
at x=0\displaystyle \mx{\text{at } x = 0}
absolute minimum value −6,\displaystyle \mn{\text{absolute minimum value } {-6},}
at x=−1\displaystyle \mn{\text{at } x = -1}
5
Find the absolute maximum and minimum values of f(x)=sin⁡x+cos⁡xf(x) = \sin x + \cos x on the interval [0,2π][0, 2\pi].

On the interval [0,2π][0, 2\pi], how many angles have cos⁡x=sin⁡x\cos x = \sin x?

Set the derivative equal to 00 to find the critical points.
f′(x)=cos⁡x−sin⁡x=0  →  cos⁡x=sin⁡x\displaystyle \slp{f'(x) = \cos x - \sin x = 0} \;\to\; \slp{\cos x = \sin x}
On the unit circle, the x-coordinate equals the y-coordinate along the line y=xy = x, in the first and third quadrants.
0≤x≤2π:x=π4,x=5π4\displaystyle \iv{0 \le x \le 2\pi}: \qquad \slp{x = \frac{\pi}{4}}, \qquad \slp{x = \frac{5\pi}{4}}
Evaluate ff at the two critical points and at the endpoints, then compare the values.
f(0)\displaystyle f(0) =1,f ⁣(π4)=2,f ⁣(5π4)=−2,f(2π)\displaystyle = 1, \qquad \mx{f\!\left(\frac{\pi}{4}\right) = \sqrt2}, \qquad \mn{f\!\left(\frac{5\pi}{4}\right) = -\sqrt2}, \qquad f(2\pi) =1\displaystyle = 1
absolute maximum value 2,\displaystyle \mx{\text{absolute maximum value } \sqrt2,}
at x=π4\displaystyle \mx{\text{at } x = \frac{\pi}{4}}
absolute minimum value −2,\displaystyle \mn{\text{absolute minimum value } {-\sqrt2},}
at x=5π4\displaystyle \mn{\text{at } x = \frac{5\pi}{4}}
6
The function ff is differentiable on [0,6][0, 6], and f′(x)=0f'(x) = 0 only at x=1x = 1 and x=4x = 4. The table gives f(0)=2f(0) = 2, f(1)=5f(1) = 5, f(2)=4f(2) = 4, f(4)=−3f(4) = -3, f(6)=1f(6) = 1. Find the absolute maximum and minimum values of ff on [0,6][0, 6]. Justify your answer.

Since ff is differentiable on [0,6][0, 6], can f′f' fail to exist anywhere? What does that leave for candidates?

Because ff is differentiable, f′\slp{f'} exists everywhere, so the only candidates are the endpoints and the points where f′(x)=0\slp{f'(x) = 0}.
f differentiable on [0,6]  →  f continuous,f′(x) exists for every x\displaystyle f \text{ differentiable on } \iv{[0, 6]} \;\to\; f \text{ continuous}, \quad \slp{f'(x) \text{ exists for every } x}
candidates: x\displaystyle \text{candidates: } x =0, 1, 4, 6\displaystyle = 0, \ 1, \ 4, \ 6
Since x=2x = 2 isn't an endpoint or a critical point, leave it out and compare the values at the candidates.
f(0)\displaystyle f(0) =2,f(1)=5,f(4)=−3,f(6)\displaystyle = 2, \qquad \mx{f(1) = 5}, \qquad \mn{f(4) = -3}, \qquad f(6) =1\displaystyle = 1
absolute maximum value 5,\displaystyle \mx{\text{absolute maximum value } 5,}
at x=1\displaystyle \mx{\text{at } x = 1}
absolute minimum value −3,\displaystyle \mn{\text{absolute minimum value } {-3},}
at x=4\displaystyle \mn{\text{at } x = 4}
7
A particle moves along a line with position x(t)=t3−9t2+15tx(t) = t^3 - 9t^2 + 15t. Find its maximum speed for 1≤t≤51 \le t \le 5. (Challenge)

Speed is ∣v(t)∣|v(t)|. What are the largest and smallest values of v(t)v(t) on [1,5][1, 5]?

Find the velocityDifferentiate the position to get v(t)v(t), and remember that speed is ∣v(t)∣|v(t)|.
v(t)\displaystyle v(t) =x′(t)\displaystyle = x'(t) =3t2−18t+15\displaystyle = 3t^2 - 18t + 15 =3(t−1)(t−5),speed\displaystyle = 3(t-1)(t-5), \qquad \text{speed} =∣v(t)∣\displaystyle = |v(t)|
Find the extreme values of velocityDifferentiate v(t)v(t) to find its critical points, then test them along with the endpoints.
a(t)=v′(t)=6t−18=0  →  t=3\displaystyle \slp{a(t) = v'(t) = 6t - 18 = 0} \;\to\; \slp{t = 3}
v(1)\displaystyle v(1) =0,v(3)=−12,v(5)\displaystyle = 0, \qquad \mn{v(3) = -12}, \qquad v(5) =0\displaystyle = 0
Take the absolute valueThe velocity stays between −12\mn{-12} and 00, so the speed ∣v(t)∣|v(t)| is at most 12\mx{12}.
−12\displaystyle \mn{-12} ≤v(t)\displaystyle \le v(t) ≤0  →  ∣v(t)∣\displaystyle \le 0 \;\to\; |v(t)| ≤12,∣v(3)∣=12\displaystyle \le \mx{12}, \qquad \mx{|v(3)| = 12}
maximum speed 12,\displaystyle \mx{\text{maximum speed } 12,}
at t=3\displaystyle \mx{\text{at } t = 3}