A highest and a lowest value
A closed interval [a, b] contains both of its ends: every x with . The extreme value theorem says that a function continuous on a closed interval reaches a largest value and a smallest value on it, each at some x in the interval. These are the absolute maximum and absolute minimum on [a, b].
The theorem says they exist. It does not say where they are; that takes calculus.
An extreme at an end
Take on [0, 3]. Its derivative is f'(x) = 2x − 4. At the left end, f'(0) = −4, and at the right end, f'(3) = 6 − 4 = 2. Neither is 0.
Yet f(0) = 3 is the highest value on the interval. The curve is still falling at x = 0; it is highest there only because the interval starts there. Solving f'(x) = 0 cannot find a maximum like this, because the tangent at it is not level.
Every candidate
Inside the interval, a highest or lowest value is a local extreme, so it sits at a critical point, where f' is zero or does not exist. Otherwise it sits at an end. There is nowhere else.
That gives the closed interval method. Find the critical points of f that lie strictly between a and b. Work out f at each of them, and at a and b. The largest of these values is the absolute maximum and the smallest is the absolute minimum.
No test is needed to classify the critical points: comparing the values does that.
Comparing the values
For on [0, 3], f'(x) = 2x − 4 = 0 gives x = 2, which lies inside the interval. The candidates are x = 0, x = 2 and x = 3.
f(0) = 3. f(2) = 4 − 8 + 3 = −1. f(3) = 9 − 12 + 3 = 0.
The largest value is 3, so the absolute maximum is 3, at the end x = 0. The smallest is −1, so the absolute minimum is −1, at the critical point x = 2. The value 0 at the other end is neither.
The gold curve , drawn from x = 0 to x = 3, with the three candidates marked. The highest is the left end, (0, 3), where the curve is still falling; the lowest is the trough, (2, −1).
The same curve, different intervals
Take . Its derivative is zero at x = −1 and x = 1.
On [0, 3], only x = 1 lies inside. The values are f(0) = 0, f(1) = 1 − 3 = −2 and f(3) = 27 − 9 = 18. The absolute minimum is −2 at x = 1, and the absolute maximum is 18 at the end x = 3.
On [−2, 2], both critical points lie inside. The values are f(−2) = −8 + 6 = −2, f(−1) = −1 + 3 = 2, f(1) = −2 and f(2) = 8 − 6 = 2. The absolute maximum is 2, reached twice, at x = −1 and at x = 2; the absolute minimum is −2, reached at x = −2 and at x = 1. An absolute maximum is one value, but it can be taken at more than one x.
The gold curve on [−2, 2]. The gold line y = 2 touches it at the peak x = −1 and at the end x = 2; the dashed line y = −2 at the end x = −2 and the trough x = 1.
A critical point with no derivative
Take on [−1, 8]. Its derivative is never 0, and does not exist at x = 0, which is inside the interval. So the candidates are x = −1, x = 0 and x = 8.
f(−1) = 1, since the cube root of −1 is −1 and its square is 1. f(0) = 0. . The absolute maximum is 4 at the end x = 8, and the absolute minimum is 0 at the cusp x = 0. Solving f' = 0 alone would have found no candidate inside at all.
When the theorem does not apply
Take on (0, 2], the interval with the end 0 left out. Close to 0 the values grow without bound: f(0.01) = 100, f(0.0001) = 10000. There is no largest value, because whatever value is reached, a smaller x gives a larger one. The smallest value, at x = 2, does exist.
Continuity matters as much as the ends. On the closed interval [−1, 1], is not continuous, since it has no value at 0, and it has neither a maximum nor a minimum there: it runs off upwards just right of 0 and downwards just left of it.
So check both conditions, closed and continuous, before promising that a maximum exists.
The gold curve on (0, 2]. It reaches its lowest value, , at the end x = 2, which is included; toward the missing end at 0 it climbs past every height.
The usual mistakes
Comparing only the critical points. For on [0, 3], the one critical point gives −1, and the maximum, 3, is at an end.
Comparing only the ends. For on [0, 3], the ends give 0 and 18, and the minimum, −2, is inside at x = 1.
Including a critical point outside the interval. For on [0, 3], x = −1 is not in the interval, and its value 2 is not a candidate.
Giving x when the value is asked for. The absolute maximum of on [0, 3] is 3; it occurs at x = 0.
A delivery van between two speeds
In the application below, a van must keep between 60 and 120 km/h. The fuel it uses has one stationary point inside that range, and the greatest use is at an end.
Worked example: A Delivery Van Held Between Two Speeds: the Best Speed and the Worst in the Allowed Range
Question A delivery van uses F = 14 − 0.2v + 0.00125v2 liters of fuel per 100 km when it is driven at a steady v km/h. On its route the van may not go slower than 60 km/h nor faster than 120 km/h. (a) Find the speed that uses least fuel, and how much it uses. (b) Find the speed that uses most fuel over the allowed range, and how much it uses there.
1.Differentiate: dFdv = −0.2 + 0.0025v liters per 100 km for each km/h.
Differentiating the fuel used gives dFdv = −0.2 + 0.0025v. 2.Set it to zero for the stationary point: −0.2 + 0.0025v = 0, so v = 80, which lies inside the allowed range.
Setting it to zero gives v = 80 km/h, which lies inside the allowed range. 3.(a) At v = 80, F = 14 − 16 + 0.00125 × 6400 = 14 − 16 + 8 = 6 liters per 100 km. Here d2Fdv2 = 0.0025 is positive, so this stationary point is a minimum, and it is the least fuel the van can use.
(a) There F = 14 − 16 + 8 = 6 liters per 100 km, and d2Fdv2 = 0.0025 is positive, so it is the least. 4.Now work out both ends. At v = 60, F = 14 − 12 + 0.00125 × 3600 = 14 − 12 + 4.5 = 6.5; at v = 120, F = 14 − 24 + 0.00125 × 14400 = 14 − 24 + 18 = 8.
Both ends must be worked out too: F = 6.5 at 60 km/h and F = 8 at 120 km/h. 5.Compare the three values: 6 liters at 80 km/h, 6.5 liters at 60 km/h and 8 liters at 120 km/h.
Comparing the three: 6 liters at 80, 6.5 at 60 and 8 at 120. 6.(b) The most fuel is used at 120 km/h, at 8 liters per 100 km. The gradient there is −0.2 + 0.3 = 0.1, which is not zero: the interval simply stops before the curve does, so the greatest value sits at an end.
(b) The most fuel is used at 120 km/h, at 8 liters per 100 km, where the gradient is 0.1 and not zero.
Answer: (a) 80 km/h, using 6 liters per 100 km; (b) 120 km/h, using 8 liters per 100 km
Common mistakes
- Solving dFdv = 0 and offering that one speed as both the best and the worst. A stationary point gives only a local turning value; on a closed interval the two ends have to be compared with it, and here the greatest value is at an end.
- Reading F as the fuel for the whole journey. It is liters per 100 km, so a longer route uses more fuel at every speed, while the speed that uses least stays 80 km/h.