Class 12 Maths, Worked in Full Past papersGet updates
Chapter 6 · Application of Derivatives

Exercise 6.3 · Q4: Why exe^x⁠, log⁡x\log x and x3+x2+x+1x^3+x^2+x+1 have no maxima or minima

Exercise 6.3Critical points · Theorem 2 · Section 6.4
The question

4. Prove that the following functions do not have maxima or minima:

(i) f(x)=exf(x)=e^x    (ii) g(x)=log⁡xg(x)=\log x    (iii) h(x)=x3+x2+x+1h(x)=x^3+x^2+x+1

Exam answer

At a point of local maxima or minima of a differentiable function, its derivative is 00⁠.

(i) f′(x)=ex>0f'(x) = e^x > 0 for all xx⁠, so f′(x)≠0f'(x) \ne 0 at every point.

So ff has no maxima or minima.

(ii) On (0,∞)(0, \infty)⁠, g′(x)=1x>0g'(x) = \frac{1}{x} > 0⁠, so g′(x)≠0g'(x) \ne 0 at every point.

So gg has no maxima or minima.

(iii) h′(x)=3x2+2x+1h'(x) = 3x^2 + 2x + 1⁠. Its discriminant is 22−4⋅3⋅1=−8<02^2 - 4\cdot3\cdot1 = -8 < 0 and its leading coefficient is positive, so h′(x)>0h'(x) > 0 for all xx⁠.

So hh has no maxima or minima.

Want the reason behind every step, the basics and the checks?

The one idea this problem is built on

Theorem 2 says: at a point of local maxima or minima inside an open interval, the derivative is zero or does not exist. All three functions are differentiable everywhere on their domains, and their derivatives are never zero (in fact always positive). So there is no point at which a maximum or minimum could sit. A maximum or minimum value over the whole domain would also be a local one, because the domain is open, so those are ruled out too.

Given, to find, and the plan
Given
f(x)=exf(x)=e^x on R\R⁠; g(x)=log⁡xg(x)=\log x on (0,∞)(0,\infty) (log⁡\log is the natural logarithm, defined only for positive xx⁠); h(x)=x3+x2+x+1h(x)=x^3+x^2+x+1 on R\R⁠.
To show
None of them has a point of local maxima or local minima (Definition 4), and so none has a maximum or minimum value (Definition 3).
Plan
(1) Prove a general link: on an open interval, a point of maximum or minimum value is also a point of local maxima or minima. (2) For each function, show the derivative exists everywhere and is never zero. (3) Conclude with Theorem 2. (4) Confirm directly by beating every candidate point.

Basics you need

1Theorem 2

Let ff be defined on an open interval I and c∈c\in I. If ff has a local maxima or a local minima at x=cx=c⁠, then either f′(c)=0f'(c)=0 or ff is not differentiable at cc⁠. Lesson: critical points and tests

2Local and overall extremes

Local maxima (Definition 4): f(c)≥f(x)f(c)\ge f(x) for all xx near cc⁠. Maximum value (Definition 3): f(c)≥f(x)f(c)\ge f(x) for all xx in the domain. (The book prints Definition 3 with >\gt⁠, <\lt⁠; with x=cx=c in I they must be read as ≥\ge⁠, ≤\le⁠.) Lesson: maxima and minima

3Derivatives of exe^x and log⁡x\log x

ddxex=ex\frac{d}{dx}e^x=e^x for all xx⁠; ddxlog⁡x=1x\frac{d}{dx}\log x=\frac1x for x>0x\gt0⁠. Also ex>0e^x\gt0 for every xx⁠.

4A quadratic with negative discriminant

If a>0a\gt0 and b2−4ac<0b^2-4ac\lt0⁠, then ax2+bx+c>0ax^2+bx+c\gt0 for every real xx⁠. Basics: factorising and the discriminant

Part (i): f(x)=exf(x)=e^x

  1. A maximum or minimum value would be a local one

    This step serves all three parts. Suppose a function FF on an open interval I had a maximum value at some c∈c\in I: F(c)≥F(x)F(c)\ge F(x) for every x∈x\in I. Because I is open, there is an h>0h\gt0 with (c−h,c+h)(c-h,c+h) inside I, and the inequality holds in particular for those xx⁠. That is exactly Definition 4(a): cc is a point of local maxima. The same holds for minimum values.

    So once we show there are no points of local maxima or minima, there can be no maximum or minimum values either. The domains here, R\R and (0,∞)(0,\infty)⁠, are open intervals, so this applies.
    Definitions 3 and 4, Section 6.4. Lesson: maxima and minima
  2. f′(x)=exf'(x)=e^x is never zero

    ff is differentiable at every real xx⁠, with f′(x)=exf'(x)=e^x⁠. An exponential with a positive base is always positive, so f′(x)=ex>0≠0f'(x)=e^x\gt0\ne0 for every xx⁠.

    By Theorem 2, a point of local maxima or minima would need f′(c)=0f'(c)=0 or ff not differentiable at cc⁠. Neither ever happens, so exe^x has no local maxima or minima, and by Part (i), Step 1, no maximum or minimum value.

    There are no critical points at all. Theorem 2 is a necessary condition: no critical point means nowhere for an extreme to be.
    Theorem 2, Section 6.4. Lesson: tests
  3. Direct confirmation

    For any cc⁠: f(c+1)=e⋅ec>ec=f(c)f(c+1)=e\cdot e^c\gt e^c=f(c) and f(c−1)=ec/e<ecf(c-1)=e^c/e\lt e^c⁠, because e>1e\gt1⁠. Every candidate is beaten on both sides. The values exe^x get as close to 00 as we like (for example e−10≈0.0000454e^{-10}\approx0.0000454⁠) but never reach it, just as in Example 16.

    This shows the conclusion without Theorem 2, which is a good check that we read the theorem in the right direction.

Part (ii): g(x)=log⁡xg(x)=\log x

  1. g′(x)=1xg'(x)=\frac1x is never zero on (0,∞)(0,\infty)

    The domain of log⁡x\log x is (0,∞)(0,\infty)⁠, an open interval. On it gg is differentiable with g′(x)=1x>0g'(x)=\frac1x\gt0⁠. A fraction with numerator 11 is never zero. So there is no critical point, and by Theorem 2 log⁡x\log x has no local maxima or minima; by Part (i), Step 1, no maximum or minimum value.

    The point x=0x=0 might look special, but it is not in the domain, so it cannot be a point of maxima or minima.
    ddxlog⁡x=1x\frac{d}{dx}\log x=\frac1x⁠, Chapter 5.
  2. Direct confirmation

    For any c>0c\gt0⁠: log⁡(2c)=log⁡2+log⁡c>log⁡c\log(2c)=\log2+\log c\gt\log c and log⁡c2=log⁡c−log⁡2<log⁡c\log\frac{c}{2}=\log c-\log2\lt\log c⁠, since log⁡2>0\log2\gt0⁠. Both 2c2c and c2\frac c2 are positive, so they are in the domain.

    Doubling or halving keeps us inside (0,∞)(0,\infty)⁠, unlike adding or subtracting 11⁠, which could leave the domain when cc is small.
    log⁡(ab)=log⁡a+log⁡b\log(ab)=\log a+\log b⁠.
x y O −2 1 2 4 −2 1 2 4 y = eˣ y = log x
y=exy=e^x and y=log⁡xy=\log x (mirror images in the dashed line y=xy=x⁠). Both rise all the way across; neither has a hilltop or a valley.

Part (iii): h(x)=x3+x2+x+1h(x)=x^3+x^2+x+1

  1. h′(x)=3x2+2x+1h'(x)=3x^2+2x+1 is always positive

    hh is a polynomial, differentiable everywhere, with h′(x)=3x2+2x+1h'(x)=3x^2+2x+1⁠. Its discriminant is 22−4⋅3⋅1=−8<02^2-4\cdot3\cdot1=-8\lt0 and its leading coefficient 33 is positive, so h′(x)>0h'(x)\gt0 for every xx⁠. Without the discriminant:

    3x2+2x+1=2x2+(x+1)2.3x^2+2x+1=2x^2+(x+1)^2.

    Both terms are ≥0\ge0⁠, and they cannot both be zero (that would need x=0x=0 and x=−1x=-1 at once), so the sum is >0\gt0⁠.

    Expanding: 2x2+(x2+2x+1)=3x2+2x+12x^2+(x^2+2x+1)=3x^2+2x+1⁠. ✓ Completing the square gives the exact least value: 3(x+13)2+23≥233\left(x+\frac13\right)^2+\frac23\ge\frac23⁠.
    Negative discriminant ⇒ no real roots. Basics: the discriminant
  2. Conclude with Theorem 2

    h′(c)≠0h'(c)\ne0 for every cc⁠, and hh is differentiable everywhere, so there is no critical point. By Theorem 2, hh has no local maxima or minima, and by Part (i), Step 1, no maximum or minimum value on R\R⁠.

    Since h′>0h'\gt0 everywhere, hh is in fact strictly increasing on R\R (Theorem 1 and the Mean Value step of its proof, applied on [x1,x2][x_1,x_2] for any x1<x2x_1\lt x_2⁠), so every point is beaten by any point to its right and undercuts any point to its left.
    Theorem 2, Section 6.4; Theorem 1, Section 6.3. Lesson: increasing and decreasing functions
x y O −2 −1 1 −4 4 y = h(x) y = h′(x)
h(x)=x3+x2+x+1h(x)=x^3+x^2+x+1 (solid) rises everywhere. Its derivative h′(x)=3x2+2x+1h'(x)=3x^2+2x+1 (dashed) never reaches the xx-axis: its lowest value is 23\frac23⁠, at x=−13x=-\frac13 (dot).
Result

(i) f′(x)=ex>0f'(x)=e^x\gt0⁠, (ii) g′(x)=1x>0g'(x)=\frac1x\gt0 on (0,∞)(0,\infty)⁠, (iii) h′(x)=3x2+2x+1>0h'(x)=3x^2+2x+1\gt0⁠. In each case the function is differentiable on an open interval and its derivative is never zero, so by Theorem 2 it has no point of local maxima or minima, and therefore no maximum or minimum value. Hence none of the three functions has maxima or minima. ∎

Check it with numbers

(i) e−1≈0.3679e^{-1}\approx0.3679⁠, e0=1e^0=1⁠, e1≈2.7183e^1\approx2.7183⁠: always rising, always positive.

(ii) log⁡0.5≈−0.6931\log0.5\approx-0.6931⁠, log⁡1=0\log1=0⁠, log⁡2≈0.6931\log2\approx0.6931⁠: always rising.

(iii) h′(−13)=13−23+1=23h'(-\frac13)=\frac13-\frac23+1=\frac23⁠, the smallest value of h′h'⁠; and h(−2)=−5h(-2)=-5⁠, h(−1)=0h(-1)=0⁠, h(0)=1h(0)=1⁠, h(1)=4h(1)=4⁠: always rising.

Common mistakes
  1. Reading Theorem 2 backwards. It says "extreme ⇒ critical point". We use its contrapositive: "no critical point ⇒ no extreme". It does not say every critical point is an extreme.
  2. Forgetting the domain of log⁡x\log x⁠. It is (0,∞)(0,\infty)⁠; x=0x=0 and negative xx are not candidates.
  3. Saying exe^x has minimum value 00⁠. exe^x gets arbitrarily close to 00 but never equals it.
  4. Trying to solve 3x2+2x+1=03x^2+2x+1=0 and stopping at "no solution". Say clearly what that means: no critical points, hence no local extremes.
What to take away
  1. No critical points on an open interval ⇒ no local maxima or minima (Theorem 2).
  2. On an open interval, a maximum or minimum value is automatically a local one; so no local extremes ⇒ no extreme values.
  3. To show a quadratic derivative is never zero, use the discriminant or write it as a sum of squares.

Every fact used, and where it comes from

FactUsed inSource
Maximum and minimum value; local maxima and minimaPart (i), Step 1Definitions 3 and 4, Section 6.4 (Lesson: maxima and minima)
Local extremum ⇒ f′(c)=0f'(c)=0 or not differentiableEvery partTheorem 2, Section 6.4 (Lesson: tests)
(ex)′=ex>0(e^x)'=e^x\gt0⁠; (log⁡x)′=1x(\log x)'=\frac1x⁠; power rulePart (i), Step 2; Parts (ii), (iii), Step 1Chapter 5 (Continuity and Differentiability)
log⁡(ab)=log⁡a+log⁡b\log(ab)=\log a+\log bPart (ii), Step 2Laws of logarithms (Chapter 5)
Negative discriminant and positive leading coefficient ⇒ quadratic always positivePart (iii), Step 1Class XI algebra (Basics)
f′>0f'\gt0 on an interval ⇒ ff strictly increasingPart (iii), Step 2Theorem 1 and its proof, Section 6.3
Values approached but never reached give no extreme valuePart (i), Step 3Example 16, Section 6.4
Found an error, or want more?

Tell us if something on this page is wrong or unclear, or if you would like a fuller explanation, another method or a new feature. Send feedback on this problem →