MATH 203 Calculus I • Concordia University, Montreal

Corrected exercises: extreme values of functions (MATH 203)

This is the corrected exercise set for extreme values of functions in MATH 203, Differential and Integral Calculus I, at Concordia University, section 4.1 of Thomas' Calculus. It opens the applications of the derivative: before a curve is sketched or a quantity optimized, the candidates for an extremum have to be found and compared. The definitions are those of Thomas, and two of them differ from other textbooks: a critical point is an INTERIOR point of the domain, and a local extremum can sit at an endpoint.

The thread running through the whole set: the candidates hide in the algebra of f′f'. Thomas's Theorem 2 says where to look, interior points where f′=0f' = 0 or f′f' is undefined, but both kinds only appear once f′f' is written as ONE factored fraction: the lowest power factored out of x−1/5x^{-1/5} and x4/5x^{4/5}, a square root put over a common denominator, sin⁡2x=1−cos⁡2x\sin^2 x = 1 - \cos^2 x applied, a common factor kept instead of divided out. After that, the scientific calculator only evaluates a finite list, and every solution names the algebraic step where the marks are usually lost.

The traps named in the solutions: dividing an equation by sin⁡x\sin x or by xx and losing a critical point, forgetting the points where f′f' is undefined, calling an endpoint a critical point or refusing it a local extremum, keeping a critical point outside the interval, evaluating f′f' instead of ff, a calculator that reads −24-2^4 as −16-16, a calculator ERROR on (−8)2/3(-8)^{2/3} taken for an undefined value, degree mode on arctan⁡3\arctan 3, applying the Extreme Value Theorem when the interval is open, and reporting a maximum that is approached but never reached.

10 corrected exercises • 100 points • 150 minutes

Revision sheet for this chapter → Every MATH 203 chapter →

Self-checking set Type your answers under each question: the page tells you right or wrong before you open the solution. With an account, every answer right on the first try earns points.

Course recap

  • • Absolute maximum of ff on DD at cc: f(x)≤f(c)f(x) \le f(c) for all x∈Dx \in D. Local maximum at cc: f(x)≤f(c)f(x) \le f(c) for all x∈Dx \in D in some open interval containing cc; in Thomas an endpoint can qualify.
  • • Theorem 1 (Extreme Value Theorem): ff continuous on a closed interval [a,b][a, b] attains an absolute maximum MM and an absolute minimum mm in [a,b][a, b].
  • • Theorem 2 (First Derivative Theorem for Local Extreme Values): local extremum at an interior point cc and f′(c)f'(c) defined imply f′(c)=0f'(c) = 0. The converse is false.
  • • Critical point: an interior point of the domain where f′=0f' = 0 or f′f' is undefined. Endpoints are never critical points.
  • • Closed interval method: check continuity on [a,b][a, b]; evaluate ff at the critical points and at the endpoints; the largest value is the absolute maximum, the smallest the absolute minimum.
  • • Algebra first: f′f' as one factored fraction. Calculator: radian mode, brackets around a negative base, (−8)2/3=((−8)1/3)2=4(-8)^{2/3} = \left((-8)^{1/3}\right)^2 = 4.

Part A: the basics (/50)

Exercise 1: Reading extrema on a graph with Thomas's definitions: endpoints included

Definitions of Thomas, section 4.1. ff has an absolute maximum value on its domain DD at cc if f(x)≤f(c)f(x) \le f(c) for ALL xx in DD. ff has a local maximum value at cc if f(x)≤f(c)f(x) \le f(c) for all xx in DD lying in some open interval containing cc. Because only the points of DD are compared, an ENDPOINT can carry a local extremum: at x=ax = a the comparison is made on a half-open interval [a,a+δ)[a, a + \delta). A critical point is an INTERIOR point of the domain where f′f' is zero or undefined.

The figure shows a function ff, continuous on [−3,4][-3, 4], with f(−3)=2f(-3) = 2, f(−2)=3f(-2) = 3, f(0)=−1f(0) = -1, f(1)=2f(1) = 2, f(2)=3f(2) = 3, f(3)=4f(3) = 4, f(4)=2f(4) = 2. It has a horizontal tangent at x=−2x = -2, a cusp at x=0x = 0, a corner at x=1x = 1 (slope 22 on the left, 13\frac{1}{3} on the right), a vertical tangent at x=2x = 2 and a corner at x=3x = 3. It is differentiable everywhere else in (−3,4)(-3, 4).

-3-2-11234-2-112345y = f(x)x
  • a) Find the absolute maximum and minimum values of ff on [−3,4][-3, 4], and where they occur.
  • b) With Thomas's definition, list the local maximum values and the local minimum values of ff, endpoints included.
  • c) List the critical points of ff. For each, say whether f′=0f' = 0 or f′f' is undefined, and whether ff has a local extremum there. Are −3-3 and 44 critical points?
  • d) Now restrict ff to [−1,2][-1, 2]. Find its absolute extrema. What is new about the point x=2x = 2?
  • e) Now restrict ff to the open interval (−3,3)(-3, 3). Does it have an absolute maximum? An absolute minimum? Does this contradict the Extreme Value Theorem?

Type your answers, the page tells you right or wrong 0/14

a)
b)
c)
d)
e)
Show the solution

Answers

  • a) Absolute max 44 at x=3x = 3; absolute min −1-1 at x=0x = 0
  • b) Local max 33 at x=−2x = -2 and 44 at x=3x = 3; local min 22 at x=−3x = -3, −1-1 at x=0x = 0, 22 at x=4x = 4
  • c) −2-2 (f′=0f' = 0); 00, 11, 22, 33 (f′f' undefined); no extremum at 11 and 22. Endpoints are not critical points.
  • d) Absolute max 33 at the endpoint x=2x = 2; absolute min −1-1 at x=0x = 0
  • e) No absolute max (values approach 44); absolute min −1-1 at x=0x = 0. No contradiction: the interval is not closed.

a) The highest point of the whole graph is the corner (3,4)(3, 4) and the lowest is the cusp (0,−1)(0, -1). Absolute maximum value 44, at x=3x = 3; absolute minimum value −1-1, at x=0x = 0. Each answer has a VALUE and a PLACE: the question asks for both, and writing max at 44 when 44 is the value costs the location mark. Every other point of the graph lies strictly between −1-1 and 44.

b) Local maximum values: f(−2)=3f(-2) = 3 (smooth top) and f(3)=4f(3) = 4 (corner). Local minimum values: f(0)=−1f(0) = -1 (cusp), and the two ENDPOINTS, f(−3)=2f(-3) = 2 and f(4)=2f(4) = 2. At x=−3x = -3 the graph rises right after the endpoint, so f(x)≥2f(x) \ge 2 on some [−3,−3+δ)[-3, -3 + \delta): in Thomas's sense this is a local minimum. The same holds at x=4x = 4, approached from the left on (4−δ,4](4 - \delta, 4]. So two local maxima and three local minima. Thomas also gives a consequence of his definition: an absolute extremum is always a local one, which is why f(3)=4f(3) = 4 and f(0)=−1f(0) = -1 appear in both lists. Some other textbooks refuse local extrema at endpoints; in MATH 203 follow Thomas, and write the word endpoint next to such an answer.

c) The critical points are INTERIOR points of [−3,4][-3, 4] where f′=0f' = 0 or f′f' is undefined. f′(−2)=0f'(-2) = 0 (horizontal tangent). f′f' is undefined at 00 (cusp), 11 and 33 (corners: two different one-sided slopes) and 22 (vertical tangent, infinite slope). Critical points: −2-2, 00, 11, 22, 33. Extrema at −2-2 (max), 00 (min), 33 (max). NONE at 11 and 22: at both, the graph keeps rising through the point, so there are values below f(c)f(c) on the left and above it on the right. Undefined f′f' nominates a point exactly like f′=0f' = 0 does, and it can be a false alarm exactly like f′=0f' = 0 can. The endpoints −3-3 and 44 are NOT critical points, since they are not interior; they still carry the local minima of b), and they are still candidates for absolute extrema.

d) On [−1,2][-1, 2] the candidates are the endpoints −1-1 and 22, and the critical points inside (−1,2)(-1, 2), namely 00 and 11. Values: f(−1)=3−1=2f(-1) = 3 - 1 = 2 (from the first piece, 3−(x+2)23 - (x + 2)^2), f(0)=−1f(0) = -1, f(1)=2f(1) = 2, f(2)=3f(2) = 3. Absolute maximum 33 at x=2x = 2, absolute minimum −1-1 at x=0x = 0. The point x=2x = 2 was a critical point WITHOUT extremum on [−3,4][-3, 4]; on [−1,2][-1, 2] it is an endpoint, and it carries the maximum. Absolute is always relative to the domain: the graph did not change, the list of candidates did.

e) On (−3,3)(-3, 3) the corner (3,4)(3, 4) is no longer in the domain. As x→3−x \to 3^- the values climb toward 44, but every value with −3<x<3-3 < x < 3 is strictly less than 44, and every value below 44 is beaten by a point closer to 33: NO absolute maximum. The absolute minimum f(0)=−1f(0) = -1 survives, because 00 is still in the domain. No contradiction: Thomas's Theorem 1 requires a CLOSED interval, and when a hypothesis fails the theorem simply says nothing.

Tick the exercises you have done or want to review: a free account, no password, keeps your ticks from one visit to the next and tells you which chapter to tackle next. Create your space, an email is enough.

Exercise 2: Critical points hide in the algebra: write f' as one factored fraction

Thomas's definition: a critical point of ff is an interior point of the domain of ff where f′=0f' = 0 or f′f' is undefined. Both halves are read on the SAME object: f′f' written as a single fraction, fully factored. The zeros of the numerator give f′=0f' = 0; the zeros of the denominator, where ff itself is defined, give f′f' undefined; the domain of ff filters both.

The algebra is where the marks go: a negative exponent left as it is, a square root not put over a common denominator, a factor divided out of an equation. For each function, give the domain, rewrite f′f' as one factored fraction, and list the critical points.

  • a) f(x)=x4/5(x−9)f(x) = x^{4/5}(x - 9).
  • b) g(x)=x2+16xg(x) = x^2 + \frac{16}{x}.
  • c) h(x)=(x+2)3−xh(x) = (x + 2)\sqrt{3 - x}.
  • d) k(x)=sin⁡2x+cos⁡xk(x) = \sin^2 x + \cos x, critical points in the open interval (0,2π)(0, 2\pi).
  • e) m(x)=x2ln⁡xm(x) = x^2 \ln x.

Type your answers, the page tells you right or wrong 0/9

a)
b)
c)
d)
e)
Show the solution

Answers

  • a) f′(x)=9(x−4)5x1/5f'(x) = \frac{9(x - 4)}{5x^{1/5}}; critical points 00 (f′f' undefined) and 44
  • b) g′(x)=2(x−2)(x2+2x+4)x2g'(x) = \frac{2(x - 2)(x^2 + 2x + 4)}{x^2}; critical point 22 only (00 not in the domain)
  • c) h′(x)=4−3x23−xh'(x) = \frac{4 - 3x}{2\sqrt{3 - x}}; critical point 43\frac{4}{3} only (33 is an endpoint of the domain)
  • d) k′(x)=sin⁡x(2cos⁡x−1)k'(x) = \sin x(2\cos x - 1); critical points π3\frac{\pi}{3}, π\pi, 5π3\frac{5\pi}{3}
  • e) m′(x)=x(2ln⁡x+1)m'(x) = x(2\ln x + 1); critical point e−1/2≈0.6065e^{-1/2} \approx 0.6065 (00 not in the domain)

a) Domain: all real numbers, since the fifth root is defined for every real number. Expand first: f(x)=x9/5−9x4/5f(x) = x^{9/5} - 9x^{4/5}, so f′(x)=95x4/5−365x−1/5f'(x) = \frac{9}{5}x^{4/5} - \frac{36}{5}x^{-1/5}. The algebraic gesture: factor out the LOWEST power, x−1/5x^{-1/5}, even though it is negative: f′(x)=95x−1/5(x−4)=9(x−4)5x1/5f'(x) = \frac{9}{5}x^{-1/5}(x - 4) = \frac{9(x - 4)}{5x^{1/5}}. Check the factoring by distributing: x−1/5⋅x=x4/5x^{-1/5} \cdot x = x^{4/5}. Numerator zero at x=4x = 4; denominator zero at x=0x = 0, where f(0)=0f(0) = 0 is defined and 00 is interior: f′(0)f'(0) is undefined and 00 IS a critical point. Critical points: 00 and 44. A student who sets 95x4/5−365x−1/5=0\frac{9}{5}x^{4/5} - \frac{36}{5}x^{-1/5} = 0 and multiplies through by x1/5x^{1/5} finds x=4x = 4 and never sees 00.

b) Domain: x≠0x \ne 0. Rewrite 16x=16x−1\frac{16}{x} = 16x^{-1} before differentiating: g′(x)=2x−16x−2=2x−16x2g'(x) = 2x - 16x^{-2} = 2x - \frac{16}{x^2}. Common denominator x2x^2: g′(x)=2x3−16x2=2(x3−8)x2=2(x−2)(x2+2x+4)x2g'(x) = \frac{2x^3 - 16}{x^2} = \frac{2(x^3 - 8)}{x^2} = \frac{2(x - 2)(x^2 + 2x + 4)}{x^2}, by the difference of cubes. The quadratic factor has discriminant 4−16<04 - 16 < 0, so it never vanishes. Numerator zero: x=2x = 2. Denominator zero: x=0x = 0, NOT in the domain of gg, so not a critical point. Critical point: 22 only. The frequent error is −16x−2-16x^{-2} written as 16x−216x^{-2} or as −16x-\frac{16}{x}: the power rule on x−1x^{-1} gives −x−2-x^{-2}.

c) Domain: 3−x≥03 - x \ge 0, that is (−∞,3](-\infty, 3]. Product rule and chain rule (inner function 3−x3 - x): h′(x)=3−x+(x+2)⋅−123−xh'(x) = \sqrt{3 - x} + (x + 2) \cdot \frac{-1}{2\sqrt{3 - x}}. Common denominator 23−x2\sqrt{3 - x}: multiply the first term by 23−x23−x\frac{2\sqrt{3 - x}}{2\sqrt{3 - x}}, which turns 3−x\sqrt{3 - x} into 2(3−x)23−x\frac{2(3 - x)}{2\sqrt{3 - x}}. So h′(x)=2(3−x)−(x+2)23−x=4−3x23−xh'(x) = \frac{2(3 - x) - (x + 2)}{2\sqrt{3 - x}} = \frac{4 - 3x}{2\sqrt{3 - x}} for x<3x < 3. Numerator zero: x=43x = \frac{4}{3}, interior. Denominator zero: x=3x = 3, where h(3)=0h(3) = 0 is defined, but 33 is an ENDPOINT of the domain, not an interior point: with Thomas's definition it is not a critical point. Critical point: 43\frac{4}{3} only. The point 33 is not forgotten for all that: on any closed interval ending at 33 it is a candidate as an endpoint.

d) Chain rule on sin⁡2x=(sin⁡x)2\sin^2 x = (\sin x)^2: k′(x)=2sin⁡xcos⁡x−sin⁡xk'(x) = 2\sin x\cos x - \sin x. Factor sin⁡x\sin x: k′(x)=sin⁡x(2cos⁡x−1)k'(x) = \sin x(2\cos x - 1). It exists everywhere, so only the zeros count. sin⁡x=0\sin x = 0 in (0,2π)(0, 2\pi): x=πx = \pi (00 and 2π2\pi are excluded). cos⁡x=12\cos x = \frac{1}{2} in (0,2π)(0, 2\pi): x=π3x = \frac{\pi}{3} and x=5π3x = \frac{5\pi}{3}. Critical points: π3\frac{\pi}{3}, π\pi, 5π3\frac{5\pi}{3}. The costly move is to write 2sin⁡xcos⁡x=sin⁡x2\sin x\cos x = \sin x and DIVIDE by sin⁡x\sin x: it leaves cos⁡x=12\cos x = \frac{1}{2} and silently erases x=πx = \pi. Never divide an equation by a factor that can be zero; factor it out and set each factor to zero.

e) Domain: x>0x > 0. Product rule: m′(x)=2xln⁡x+x2⋅1x=2xln⁡x+x=x(2ln⁡x+1)m'(x) = 2x\ln x + x^2 \cdot \frac{1}{x} = 2x\ln x + x = x(2\ln x + 1). The factor xx vanishes at 00, which is NOT in the domain: rejected. 2ln⁡x+1=02\ln x + 1 = 0 gives ln⁡x=−12\ln x = -\frac{1}{2}, so x=e−1/2=1e≈0.6065x = e^{-1/2} = \frac{1}{\sqrt e} \approx 0.6065. Critical point: e−1/2e^{-1/2} only. Write the exact value first; the calculator gives the decimal only after the equation is solved by hand.

Exercise 3: The Extreme Value Theorem: check the hypotheses, then see what survives

Thomas, Theorem 1 (the Extreme Value Theorem). If ff is continuous on a closed interval [a,b][a, b], then ff attains both an absolute maximum value MM and an absolute minimum value mm in [a,b][a, b].

Two hypotheses, CONTINUOUS and FINITE CLOSED interval, and one conclusion: the extrema EXIST. The theorem does not say where they are, and when a hypothesis fails it says nothing at all. For each case, name the hypothesis that fails (if any), then find the absolute extrema that exist. The figure shows the function kk of part c).

0.511.522.512345y = k(x)x
  • a) f(x)=xf(x) = \sqrt{x} on [0,4][0, 4].
  • b) f(x)=tan⁡xf(x) = \tan x on [−π4,π2)\left[-\frac{\pi}{4}, \frac{\pi}{2}\right).
  • c) k(x)=4−x22−xk(x) = \frac{4 - x^2}{2 - x} for 0≤x<20 \le x < 2, and k(2)=1k(2) = 1, on [0,2][0, 2].
  • d) f(x)=6xx2+9f(x) = \frac{6x}{x^2 + 9} on (−∞,∞)(-\infty, \infty). Hint: compute 1−f(x)1 - f(x) as one fraction.
  • e) Without computing anything, does g(x)=x5−3x3+excos⁡xg(x) = x^5 - 3x^3 + e^x\cos x have an absolute maximum on [−2,3][-2, 3]?

Type your answers, the page tells you right or wrong 0/13

a)
b)
c)
d)
e)
Show the solution

Answers

  • a) Hypotheses hold; max 22 at x=4x = 4, min 00 at x=0x = 0
  • b) Not closed; no absolute max, absolute min −1-1 at x=−π4x = -\frac{\pi}{4}
  • c) Not continuous at 22; no absolute max, absolute min 11 at x=2x = 2
  • d) Not a finite closed interval, yet max 11 at x=3x = 3 and min −1-1 at x=−3x = -3
  • e) Yes: gg is continuous on the closed interval [−2,3][-2, 3], so the EVT guarantees it.

a) x\sqrt x is continuous on its domain [0,∞)[0, \infty), hence on [0,4][0, 4] (at 00, continuity from the right is what the closed interval requires). The EVT applies. f′(x)=12xf'(x) = \frac{1}{2\sqrt x} is never 00, and it is undefined only at 00, an endpoint: no critical point in (0,4)(0, 4). Candidates: the endpoints. f(0)=0f(0) = 0, f(4)=2f(4) = 2. Absolute maximum 22 at x=4x = 4, absolute minimum 00 at x=0x = 0. The vertical tangent at 00 does not make 00 a critical point, since it is not interior, but 00 is on the list anyway as an endpoint.

b) tan⁡x=sin⁡xcos⁡x\tan x = \frac{\sin x}{\cos x} is continuous wherever cos⁡x≠0\cos x \ne 0, which holds on [−π4,π2)\left[-\frac{\pi}{4}, \frac{\pi}{2}\right); but the interval is NOT closed, the right end π2\frac{\pi}{2} is missing. The EVT does not apply. As x→π2−x \to \frac{\pi}{2}^-, sin⁡x→1\sin x \to 1 and cos⁡x→0+\cos x \to 0^+, so tan⁡x→+∞\tan x \to +\infty: the values are unbounded above, NO absolute maximum. The tangent is increasing on (−π2,π2)\left(-\frac{\pi}{2}, \frac{\pi}{2}\right) (a property of its graph, section 1.3), so its smallest value on the interval is at the left end: absolute minimum tan⁡(−π4)=−1\tan\left(-\frac{\pi}{4}\right) = -1, at x=−π4x = -\frac{\pi}{4}.

c) First the algebra: for x≠2x \ne 2, 4−x22−x=(2−x)(2+x)2−x=2+x\frac{4 - x^2}{2 - x} = \frac{(2 - x)(2 + x)}{2 - x} = 2 + x. So on [0,2)[0, 2), k(x)=2+xk(x) = 2 + x, whose values fill [2,4)[2, 4), and k(2)=1k(2) = 1. The interval [0,2][0, 2] is closed, but lim⁡x→2−k(x)=4≠1=k(2)\lim_{x\to 2^-} k(x) = 4 \ne 1 = k(2): kk is NOT continuous at 22, and the EVT does not apply. The value 44 is approached (hollow dot on the figure) and never taken: NO absolute maximum. The lowest value is k(2)=1k(2) = 1, below every value 2+x≥22 + x \ge 2: absolute minimum 11 at x=2x = 2 (full dot). Without the factorization, the formula looks undefined at 22 and the whole question is misread.

d) ff is continuous on (−∞,∞)(-\infty, \infty), since x2+9>0x^2 + 9 > 0, but the interval is not a finite closed interval: the EVT is silent. Follow the hint: 1−6xx2+9=x2−6x+9x2+9=(x−3)2x2+9≥01 - \frac{6x}{x^2 + 9} = \frac{x^2 - 6x + 9}{x^2 + 9} = \frac{(x - 3)^2}{x^2 + 9} \ge 0, with equality only at x=3x = 3. So f(x)≤1=f(3)f(x) \le 1 = f(3) for every xx: absolute maximum 11 at x=3x = 3. In the same way, f(x)+1=x2+6x+9x2+9=(x+3)2x2+9≥0f(x) + 1 = \frac{x^2 + 6x + 9}{x^2 + 9} = \frac{(x + 3)^2}{x^2 + 9} \ge 0: absolute minimum −1-1 at x=−3x = -3. The algebraic gesture is to recognize the perfect square x2−6x+9x^2 - 6x + 9 in the numerator. Both extrema exist although the theorem does not apply: its hypotheses are SUFFICIENT, not necessary.

e) Yes. x5−3x3x^5 - 3x^3 is a polynomial, exe^x and cos⁡x\cos x are continuous everywhere, and sums and products of continuous functions are continuous: gg is continuous on the closed interval [−2,3][-2, 3]. By the Extreme Value Theorem, gg attains an absolute maximum (and an absolute minimum) there. The theorem is an EXISTENCE result: it answers this question in one sentence, without a derivative, and without saying where the maximum is.

Exercise 4: The closed interval method on polynomials and a quotient, and a calculator that lies

Thomas's method for the absolute extrema of a function ff CONTINUOUS on a finite CLOSED interval: evaluate ff at all critical points and endpoints, then take the largest and smallest of these values.

Write the method in four lines: (1) continuity on [a,b][a, b], with its reason; (2) f′f' factored, critical points in (a,b)(a, b), the others written rejected; (3) a table of values of ff, not of f′f'; (4) the conclusion, value AND location.

  • a) f(x)=2x3−9x2+12x−1f(x) = 2x^3 - 9x^2 + 12x - 1 on [0,3][0, 3].
  • b) g(x)=x4−8x2+3g(x) = x^4 - 8x^2 + 3 on [−1,3][-1, 3].
  • c) h(x)=xx2+4h(x) = \frac{x}{x^2 + 4} on [0,5][0, 5].
  • d) A student finds the absolute extrema of p(x)=x4−2x2p(x) = x^4 - 2x^2 on [−2,1][-2, 1]. She gets p′(x)=4x(x−1)(x+1)p'(x) = 4x(x - 1)(x + 1), lists the candidates −2-2, −1-1, 00, 11, types each value on her calculator and gets p(−2)=−24p(-2) = -24, p(−1)=−1p(-1) = -1, p(0)=0p(0) = 0, p(1)=−1p(1) = -1. She concludes: absolute maximum 00 at x=0x = 0, absolute minimum −24-24 at x=−2x = -2. Find the error and give the correct answer.

Type your answers, the page tells you right or wrong 0/12

a)
b)
c)
d)
Show the solution

Answers

  • a) Max 88 at x=3x = 3; min −1-1 at x=0x = 0
  • b) Max 1212 at x=3x = 3; min −13-13 at x=2x = 2 (x=−2x = -2 rejected)
  • c) Max 14\frac{1}{4} at x=2x = 2; min 00 at x=0x = 0 (x=−2x = -2 rejected)
  • d) −24-2^4 was read as −(24)-(2^4): p(−2)=16−8=8p(-2) = 16 - 8 = 8. Max 88 at x=−2x = -2; min −1-1 at x=−1x = -1 and x=1x = 1

a) ff is a polynomial, continuous on [0,3][0, 3]. f′(x)=6x2−18x+12=6(x2−3x+2)=6(x−1)(x−2)f'(x) = 6x^2 - 18x + 12 = 6(x^2 - 3x + 2) = 6(x - 1)(x - 2): critical points 11 and 22, both in (0,3)(0, 3). Values: f(0)=−1f(0) = -1, f(1)=2−9+12−1=4f(1) = 2 - 9 + 12 - 1 = 4, f(2)=16−36+24−1=3f(2) = 16 - 36 + 24 - 1 = 3, f(3)=54−81+36−1=8f(3) = 54 - 81 + 36 - 1 = 8. Absolute maximum 88 at x=3x = 3, absolute minimum −1-1 at x=0x = 0. Both critical points lose to the endpoints: f(1)=4f(1) = 4 is a local maximum and f(2)=3f(2) = 3 a local minimum, neither absolute. The derivative only fills the list; the values decide.

b) gg is a polynomial, continuous on [−1,3][-1, 3]. g′(x)=4x3−16x=4x(x2−4)=4x(x−2)(x+2)g'(x) = 4x^3 - 16x = 4x(x^2 - 4) = 4x(x - 2)(x + 2). Zeros 00, 22, −2-2; the last is outside (−1,3)(-1, 3): REJECTED. Values: g(−1)=1−8+3=−4g(-1) = 1 - 8 + 3 = -4, g(0)=3g(0) = 3, g(2)=16−32+3=−13g(2) = 16 - 32 + 3 = -13, g(3)=81−72+3=12g(3) = 81 - 72 + 3 = 12. Absolute maximum 1212 at x=3x = 3, absolute minimum −13-13 at x=2x = 2. Note that g(−2)=−13g(-2) = -13 too: keeping −2-2 gives the right minimum VALUE at a location that is not in the interval, and costs the location mark. The factoring gesture: take out the common factor 4x4x first, then the difference of squares; dividing 4x3=16x4x^3 = 16x by xx loses the critical point 00.

c) The denominator x2+4x^2 + 4 never vanishes, so hh is continuous on [0,5][0, 5]. Quotient rule: h′(x)=1⋅(x2+4)−x⋅2x(x2+4)2=4−x2(x2+4)2=(2−x)(2+x)(x2+4)2h'(x) = \frac{1 \cdot (x^2 + 4) - x \cdot 2x}{(x^2 + 4)^2} = \frac{4 - x^2}{(x^2 + 4)^2} = \frac{(2 - x)(2 + x)}{(x^2 + 4)^2}. The sign trap is in the numerator: the minus applies to ALL of x⋅2xx \cdot 2x, giving x2+4−2x2=4−x2x^2 + 4 - 2x^2 = 4 - x^2. Zeros ±2\pm 2; keep 22, reject −2-2. Values: h(0)=0h(0) = 0, h(2)=28=14h(2) = \frac{2}{8} = \frac{1}{4}, h(5)=529≈0.1724h(5) = \frac{5}{29} \approx 0.1724. Absolute maximum 14\frac{1}{4} at x=2x = 2, absolute minimum 00 at x=0x = 0.

d) The derivative and the list of candidates are correct: −1-1 and 00 are the critical points in (−2,1)(-2, 1), and −2-2, 11 are the endpoints. The error is in ONE value: the student typed -2^4 on the calculator, which computes −(24)=−16-(2^4) = -16, because the power is done before the negation. So she got −16−2⋅4=−24-16 - 2 \cdot 4 = -24. The correct value is p(−2)=(−2)4−2(−2)2=16−8=8p(-2) = (-2)^4 - 2(-2)^2 = 16 - 8 = 8: type the brackets. Correct table: p(−2)=8p(-2) = 8, p(−1)=1−2=−1p(-1) = 1 - 2 = -1, p(0)=0p(0) = 0, p(1)=−1p(1) = -1. Absolute maximum 88 at x=−2x = -2; absolute minimum −1-1, attained at x=−1x = -1 AND x=1x = 1. A five-second check would have caught it: p(x)=(x2−1)2−1≥−1p(x) = (x^2 - 1)^2 - 1 \ge -1 for every xx, so a value of −24-24 is impossible.

Exercise 5: The closed interval method with roots and fractional exponents

With a root or a fractional exponent, the critical points where f′f' is UNDEFINED appear, and they are exactly the ones a student who only solves f′(x)=0f'(x) = 0 never finds. The algebraic gesture of this exercise: rewrite f′f' as one fraction, by factoring out the lowest power or by a common denominator, before reading anything.

The calculator gesture: a scientific calculator often answers ERROR to a fractional power of a negative number, such as (−4)2/3(-4)^{2/3}. The number exists: compute it as ((−4)2)1/3=161/3\left((-4)^2\right)^{1/3} = 16^{1/3}, or as ((−4)1/3)2\left((-4)^{1/3}\right)^2. Give exact values, then decimals to 4 places.

  • a) f(x)=x2/3(5−2x)f(x) = x^{2/3}(5 - 2x) on [−1,2][-1, 2].
  • b) g(x)=x25−xg(x) = x^2\sqrt{5 - x} on [−1,5][-1, 5].
  • c) For the function gg of part b): is x=5x = 5 a critical point of gg in Thomas's sense? Is it a candidate in the closed interval method? And x=0x = 0?
  • d) h(x)=(x2−4)2/3h(x) = \left(x^2 - 4\right)^{2/3} on [−1,3][-1, 3].

Type your answers, the page tells you right or wrong 0/14

a)
b)
c)
d)
Show the solution

Answers

  • a) Max 77 at x=−1x = -1; min 00 at x=0x = 0 (where f′f' is undefined)
  • b) Max 1616 at x=4x = 4; min 00 at x=0x = 0 and x=5x = 5
  • c) 55: not a critical point (endpoint), but a candidate. 00: a critical point (g′(0)=0g'(0) = 0) and a candidate.
  • d) Max 52/3≈2.92405^{2/3} \approx 2.9240 at x=3x = 3; min 00 at x=2x = 2

a) x2/3=(x1/3)2x^{2/3} = \left(x^{1/3}\right)^2 is defined and continuous for every real xx, so ff is continuous on [−1,2][-1, 2]. Expand: f(x)=5x2/3−2x5/3f(x) = 5x^{2/3} - 2x^{5/3}, so f′(x)=103x−1/3−103x2/3f'(x) = \frac{10}{3}x^{-1/3} - \frac{10}{3}x^{2/3}. Factor out the lowest power x−1/3x^{-1/3}: f′(x)=103x−1/3(1−x)=10(1−x)3x1/3f'(x) = \frac{10}{3}x^{-1/3}(1 - x) = \frac{10(1 - x)}{3x^{1/3}}. Numerator zero at x=1x = 1; denominator zero at x=0x = 0, where f(0)=0f(0) = 0 exists: 00 is a critical point with f′f' undefined (a cusp). Values: f(−1)=(−1)2/3(5+2)=1⋅7=7f(-1) = (-1)^{2/3}(5 + 2) = 1 \cdot 7 = 7, since (−1)2/3=((−1)1/3)2=(−1)2=1(-1)^{2/3} = \left((-1)^{1/3}\right)^2 = (-1)^2 = 1; f(0)=0f(0) = 0; f(1)=1⋅3=3f(1) = 1 \cdot 3 = 3; f(2)=22/3⋅1≈1.5874f(2) = 2^{2/3} \cdot 1 \approx 1.5874. Absolute maximum 77 at x=−1x = -1, absolute minimum 00 at x=0x = 0. Forget the cusp, and the minimum becomes 1.58741.5874, which is false; trust a calculator ERROR at x=−1x = -1, and the maximum disappears.

b) 5−x≥05 - x \ge 0 on [−1,5][-1, 5], so gg is defined and continuous there. Product and chain rules: g′(x)=2x5−x+x2⋅−125−xg'(x) = 2x\sqrt{5 - x} + x^2 \cdot \frac{-1}{2\sqrt{5 - x}}. Common denominator 25−x2\sqrt{5 - x}: g′(x)=4x(5−x)−x225−x=20x−5x225−x=5x(4−x)25−xg'(x) = \frac{4x(5 - x) - x^2}{2\sqrt{5 - x}} = \frac{20x - 5x^2}{2\sqrt{5 - x}} = \frac{5x(4 - x)}{2\sqrt{5 - x}} for x<5x < 5. Zeros in (−1,5)(-1, 5): x=0x = 0 and x=4x = 4. Values: g(−1)=6≈2.4495g(-1) = \sqrt 6 \approx 2.4495, g(0)=0g(0) = 0, g(4)=16⋅1=16g(4) = 16 \cdot 1 = 16, g(5)=25⋅0=0g(5) = 25 \cdot 0 = 0. Absolute maximum 1616 at x=4x = 4; absolute minimum 00, attained TWICE, at x=0x = 0 and x=5x = 5. Report both places. The figure of the solution shows the four candidates.

c) x=5x = 5 is an endpoint of [−1,5][-1, 5] and of the domain (−∞,5](-\infty, 5] of gg. Thomas defines a critical point as an INTERIOR point where f′f' is zero or undefined, so 55 is NOT a critical point, although g′(5)g'(5) does not exist. It IS a candidate: the method evaluates ff at all critical points AND endpoints, and here 55 even carries half of the minimum. x=0x = 0 is interior with g′(0)=0g'(0) = 0: a critical point, and a candidate. In Thomas's vocabulary, the list of the method is the critical points plus the endpoints; the word critical is not stretched to cover the endpoints.

d) x2−4x^2 - 4 is a polynomial and u↦u2/3u \mapsto u^{2/3} is continuous for every real uu, so hh is continuous on [−1,3][-1, 3]. Chain rule (inner function x2−4x^2 - 4): h′(x)=23(x2−4)−1/3⋅2x=4x3(x2−4)1/3h'(x) = \frac{2}{3}\left(x^2 - 4\right)^{-1/3} \cdot 2x = \frac{4x}{3\left(x^2 - 4\right)^{1/3}}. Numerator zero: x=0x = 0. Denominator zero: x2=4x^2 = 4, x=±2x = \pm 2; keep 22, which is in (−1,3)(-1, 3), reject −2-2. At 22, h(2)=0h(2) = 0 exists: critical point with h′h' undefined. Values: h(−1)=(−3)2/3=91/3≈2.0801h(-1) = (-3)^{2/3} = 9^{1/3} \approx 2.0801; h(0)=(−4)2/3=161/3≈2.5198h(0) = (-4)^{2/3} = 16^{1/3} \approx 2.5198; h(2)=0h(2) = 0; h(3)=52/3≈2.9240h(3) = 5^{2/3} \approx 2.9240. Absolute maximum 52/3≈2.92405^{2/3} \approx 2.9240 at x=3x = 3, absolute minimum 00 at x=2x = 2. Two of the four values need the calculator gesture of the intro: typing (-4)^(2/3) often returns ERROR, and a value that the calculator refuses is still a value of hh. Here dropping x=0x = 0 and x=−1x = -1 would leave the answer unchanged by luck; in part a), dropping x=−1x = -1 loses the maximum.

-112345-224681012141618max 16 at x = 4min 0 at x = 0 and x = 5x

Part B: problems and reasoning (/50)

Exercise 6: Transcendental functions: exact candidates, then the calculator compares

With trigonometric, exponential, logarithmic and inverse trigonometric functions, the critical points are solved BY HAND (the calculator does not solve equations), and the calculator then evaluates the finite list. Set it in RADIAN mode before the first value: sin⁡π6\sin\frac{\pi}{6} must display 0.50.5.

For each function: continuity on the interval, f′f' factored, critical points inside, a table of exact values with their decimals to 4 places, the conclusion.

  • a) f(x)=sin⁡x (1+cos⁡x)f(x) = \sin x\,(1 + \cos x) on [0,5π3]\left[0, \frac{5\pi}{3}\right]. Hint: rewrite f′f' with cos⁡x\cos x only.
  • b) g(x)=xe−x2/2g(x) = x e^{-x^2/2} on [−1,2][-1, 2].
  • c) h(x)=ln⁡(x2+1)−xh(x) = \ln\left(x^2 + 1\right) - x on [0,3][0, 3].
  • d) k(x)=arctan⁡x−x2k(x) = \arctan x - \frac{x}{2} on [0,3][0, 3].
  • e) m(x)=ln⁡xxm(x) = \frac{\ln x}{\sqrt x} on [1,e3]\left[1, e^3\right].

Type your answers, the page tells you right or wrong 0/12

a)
b)
c)
d)
e)
Show the solution

Answers

  • a) Max 334≈1.2990\frac{3\sqrt 3}{4} \approx 1.2990 at x=π3x = \frac{\pi}{3}; min −334≈−1.2990-\frac{3\sqrt 3}{4} \approx -1.2990 at x=5π3x = \frac{5\pi}{3}
  • b) Max e−1/2≈0.6065e^{-1/2} \approx 0.6065 at x=1x = 1; min −e−1/2≈−0.6065-e^{-1/2} \approx -0.6065 at x=−1x = -1
  • c) Max 00 at x=0x = 0; min ln⁡10−3≈−0.6974\ln 10 - 3 \approx -0.6974 at x=3x = 3 (the critical point 11 is neither)
  • d) Max π4−12≈0.2854\frac{\pi}{4} - \frac{1}{2} \approx 0.2854 at x=1x = 1; min arctan⁡3−32≈−0.2510\arctan 3 - \frac{3}{2} \approx -0.2510 at x=3x = 3
  • e) Max 2e≈0.7358\frac{2}{e} \approx 0.7358 at x=e2x = e^2; min 00 at x=1x = 1

a) ff is a product of continuous functions. Product rule: f′(x)=cos⁡x(1+cos⁡x)+sin⁡x(−sin⁡x)=cos⁡x+cos⁡2x−sin⁡2xf'(x) = \cos x(1 + \cos x) + \sin x(-\sin x) = \cos x + \cos^2 x - \sin^2 x. The identity sin⁡2x=1−cos⁡2x\sin^2 x = 1 - \cos^2 x turns it into a quadratic in cos⁡x\cos x: f′(x)=2cos⁡2x+cos⁡x−1=(2cos⁡x−1)(cos⁡x+1)f'(x) = 2\cos^2 x + \cos x - 1 = (2\cos x - 1)(\cos x + 1). In the OPEN interval (0,5π3)\left(0, \frac{5\pi}{3}\right): cos⁡x=12\cos x = \frac{1}{2} gives x=π3x = \frac{\pi}{3} only, since the other solution 5π3\frac{5\pi}{3} is the endpoint, not an interior point; cos⁡x=−1\cos x = -1 gives x=πx = \pi. Values: f(0)=0f(0) = 0; f(π3)=32⋅32=334≈1.2990f\left(\frac{\pi}{3}\right) = \frac{\sqrt 3}{2} \cdot \frac{3}{2} = \frac{3\sqrt 3}{4} \approx 1.2990; f(π)=0f(\pi) = 0; f(5π3)=−32⋅32=−334≈−1.2990f\left(\frac{5\pi}{3}\right) = -\frac{\sqrt 3}{2} \cdot \frac{3}{2} = -\frac{3\sqrt 3}{4} \approx -1.2990. Absolute maximum 334\frac{3\sqrt 3}{4} at x=π3x = \frac{\pi}{3}, absolute minimum −334-\frac{3\sqrt 3}{4} at the endpoint 5π3\frac{5\pi}{3}, where f′f' happens to vanish too: in Thomas it is an endpoint, not a critical point, and it is on the list either way. The critical point π\pi is neither: a candidate the table dismisses.

b) gg is continuous everywhere. Product rule and chain rule (inner function −x22-\frac{x^2}{2}, derivative −x-x): g′(x)=e−x2/2+x⋅(−x)e−x2/2=e−x2/2(1−x2)g'(x) = e^{-x^2/2} + x \cdot (-x)e^{-x^2/2} = e^{-x^2/2}\left(1 - x^2\right). Factor the exponential out: it is never 00, so the zeros are those of 1−x21 - x^2, x=±1x = \pm 1. Only 11 is interior; −1-1 is an endpoint, evaluated anyway. Values: g(−1)=−e−1/2≈−0.6065g(-1) = -e^{-1/2} \approx -0.6065, g(1)=e−1/2≈0.6065g(1) = e^{-1/2} \approx 0.6065, g(2)=2e−2≈0.2707g(2) = 2e^{-2} \approx 0.2707. Absolute maximum e−1/2e^{-1/2} at x=1x = 1, absolute minimum −e−1/2-e^{-1/2} at x=−1x = -1. Writing e−x2/2=0e^{-x^2/2} = 0 as an equation to solve is a lost line: an exponential is never zero.

c) x2+1>0x^2 + 1 > 0, so hh is continuous on [0,3][0, 3]. Chain rule: h′(x)=2xx2+1−1h'(x) = \frac{2x}{x^2 + 1} - 1. Common denominator: h′(x)=2x−x2−1x2+1=−(x−1)2x2+1h'(x) = \frac{2x - x^2 - 1}{x^2 + 1} = -\frac{(x - 1)^2}{x^2 + 1}. One critical point, x=1x = 1. Values: h(0)=ln⁡1−0=0h(0) = \ln 1 - 0 = 0, h(1)=ln⁡2−1≈−0.3069h(1) = \ln 2 - 1 \approx -0.3069, h(3)=ln⁡10−3≈−0.6974h(3) = \ln 10 - 3 \approx -0.6974. Absolute maximum 00 at x=0x = 0, absolute minimum ln⁡10−3\ln 10 - 3 at x=3x = 3. The critical point 11 is neither: its value lies between the two endpoint values. The factorization shows why: h′≤0h' \le 0 on both sides of 11, a horizontal tangent in the middle of a descent.

d) kk is continuous everywhere. k′(x)=11+x2−12=2−(1+x2)2(1+x2)=1−x22(1+x2)k'(x) = \frac{1}{1 + x^2} - \frac{1}{2} = \frac{2 - (1 + x^2)}{2(1 + x^2)} = \frac{1 - x^2}{2(1 + x^2)}. Critical point in (0,3)(0, 3): x=1x = 1 (−1-1 rejected). Values: k(0)=0k(0) = 0; k(1)=π4−12≈0.2854k(1) = \frac{\pi}{4} - \frac{1}{2} \approx 0.2854; k(3)=arctan⁡3−32≈1.2490−1.5=−0.2510k(3) = \arctan 3 - \frac{3}{2} \approx 1.2490 - 1.5 = -0.2510. Absolute maximum π4−12\frac{\pi}{4} - \frac{1}{2} at x=1x = 1, absolute minimum arctan⁡3−32\arctan 3 - \frac{3}{2} at x=3x = 3. The calculator trap: in DEGREE mode, arctan⁡3\arctan 3 displays 71.565171.5651, and k(3)≈70.07k(3) \approx 70.07 becomes a false maximum. Every formula of calculus, and (arctan⁡x)′=11+x2(\arctan x)' = \frac{1}{1 + x^2} first of all, assumes radians.

e) On [1,e3]\left[1, e^3\right], x>0x > 0, so mm is continuous. Quotient rule: m′(x)=1xx−ln⁡x⋅12xxm'(x) = \frac{\frac{1}{x}\sqrt x - \ln x \cdot \frac{1}{2\sqrt x}}{x}, a complex fraction. The gesture: multiply numerator and denominator by 2x2\sqrt x. Since 1xx⋅2x=2\frac{1}{x}\sqrt x \cdot 2\sqrt x = 2, this gives m′(x)=2−ln⁡x2xx=2−ln⁡x2x3/2m'(x) = \frac{2 - \ln x}{2x\sqrt x} = \frac{2 - \ln x}{2x^{3/2}}. Zero when ln⁡x=2\ln x = 2, x=e2≈7.3891x = e^2 \approx 7.3891, which lies in (1,e3)\left(1, e^3\right). Values: m(1)=0m(1) = 0; m(e2)=2e≈0.7358m\left(e^2\right) = \frac{2}{e} \approx 0.7358; m(e3)=3e3/2≈0.6694m\left(e^3\right) = \frac{3}{e^{3/2}} \approx 0.6694. Absolute maximum 2e\frac{2}{e} at x=e2x = e^2, absolute minimum 00 at x=1x = 1.

Exercise 7: Reasoning with Thomas's two theorems: no critical point, a false alarm, constants to find

Thomas, Theorem 2 (the First Derivative Theorem for Local Extreme Values). If ff has a local maximum or minimum value at an INTERIOR point cc of its domain, and if f′f' is defined at cc, then f′(c)=0f'(c) = 0. It works in one direction only: it tells where interior extrema CAN be, never where they are.

In this chapter, a critical point is confirmed or dismissed by comparing values: by an inequality, or by the closed interval method on a small interval around it. The figure shows y=x+sin⁡xy = x + \sin x on [0,2π][0, 2\pi] and its horizontal tangent at x=πx = \pi.

12345671234567y = x + sin xflat at x = πx
  • a) Find the critical points of p(x)=x+sin⁡xp(x) = x + \sin x in (0,2π)(0, 2\pi). Show that pp has no local extremum at x=πx = \pi, using the inequality sin⁡h<h\sin h < h for h>0h > 0.
  • b) Show that r(x)=2x−cos⁡xr(x) = 2x - \cos x has no critical point at all. Deduce the absolute extrema of rr on [0,π][0, \pi] without any comparison of interior values.
  • c) Find the constants aa and bb so that f(x)=x3+ax2+bf(x) = x^3 + ax^2 + b satisfies f′(2)=0f'(2) = 0 and f(2)=5f(2) = 5. Then apply the closed interval method on [1,3][1, 3] to show that f(2)=5f(2) = 5 is a local minimum value.
  • d) Find the absolute extrema of q(x)=x2−1x2+1q(x) = \frac{x^2 - 1}{x^2 + 1} on (−∞,∞)(-\infty, \infty), if they exist. Hint: write q(x)=1−2x2+1q(x) = 1 - \frac{2}{x^2 + 1}.

Type your answers, the page tells you right or wrong 0/11

a)
b)
c)
d)
Show the solution

Answers

  • a) Only x=πx = \pi; p(π+h)>π>p(π−h)p(\pi + h) > \pi > p(\pi - h) for 0<h<π0 < h < \pi: no extremum
  • b) r′(x)=2+sin⁡x≥1r'(x) = 2 + \sin x \ge 1; on [0,π][0, \pi], min −1-1 at x=0x = 0, max 2π+1≈7.28322\pi + 1 \approx 7.2832 at x=πx = \pi
  • c) a=−3a = -3, b=9b = 9; f(1)=7f(1) = 7, f(2)=5f(2) = 5, f(3)=9f(3) = 9: min 55 on [1,3][1, 3] at an interior point, so a local minimum
  • d) Absolute min −1-1 at x=0x = 0; no absolute max (values approach 11)

a) p′(x)=1+cos⁡xp'(x) = 1 + \cos x exists everywhere and vanishes when cos⁡x=−1\cos x = -1: in (0,2π)(0, 2\pi), only x=πx = \pi. Now compare values near π\pi, with p(π)=πp(\pi) = \pi. For 0<h<π0 < h < \pi: p(π+h)=π+h+sin⁡(π+h)=π+h−sin⁡hp(\pi + h) = \pi + h + \sin(\pi + h) = \pi + h - \sin h, and since sin⁡h<h\sin h < h, this is greater than π\pi. Likewise p(π−h)=π−h+sin⁡h<πp(\pi - h) = \pi - h + \sin h < \pi. So on every open interval around π\pi there are values above and below p(π)p(\pi): NO local extremum at π\pi. The identities sin⁡(π±h)=∓sin⁡h\sin(\pi \pm h) = \mp\sin h are the algebraic step; the figure shows the curve flattening at (π,π)(\pi, \pi) and climbing on.

b) r′(x)=2+sin⁡xr'(x) = 2 + \sin x. Since sin⁡x≥−1\sin x \ge -1, r′(x)≥1>0r'(x) \ge 1 > 0 for every xx: r′r' is defined everywhere and never 00, so rr has NO critical point. By Theorem 2 read backwards (contrapositive), rr has no local extremum at any interior point. On [0,π][0, \pi], rr is continuous, the EVT guarantees the extrema, and the list of candidates reduces to the two endpoints: r(0)=0−1=−1r(0) = 0 - 1 = -1 and r(π)=2π−(−1)=2π+1≈7.2832r(\pi) = 2\pi - (-1) = 2\pi + 1 \approx 7.2832. Absolute minimum −1-1 at x=0x = 0, absolute maximum 2π+12\pi + 1 at x=πx = \pi.

c) f′(x)=3x2+2axf'(x) = 3x^2 + 2ax, so f′(2)=12+4a=0f'(2) = 12 + 4a = 0 gives a=−3a = -3. Then f(2)=8−12+b=5f(2) = 8 - 12 + b = 5 gives b=9b = 9: f(x)=x3−3x2+9f(x) = x^3 - 3x^2 + 9. These two equations only express a NECESSARY condition from Theorem 2; they do not prove an extremum. So check with values: f′(x)=3x2−6x=3x(x−2)f'(x) = 3x^2 - 6x = 3x(x - 2), whose only zero in (1,3)(1, 3) is 22. ff is continuous on [1,3][1, 3], and f(1)=1−3+9=7f(1) = 1 - 3 + 9 = 7, f(2)=5f(2) = 5, f(3)=27−27+9=9f(3) = 27 - 27 + 9 = 9. The absolute minimum of ff on [1,3][1, 3] is 55, attained at the INTERIOR point 22, so f(x)≥f(2)f(x) \ge f(2) on the open interval (1,3)(1, 3) around 22: f(2)=5f(2) = 5 is a local minimum value.

d) Long division, or adding and subtracting 11: x2−1x2+1=(x2+1)−2x2+1=1−2x2+1\frac{x^2 - 1}{x^2 + 1} = \frac{(x^2 + 1) - 2}{x^2 + 1} = 1 - \frac{2}{x^2 + 1}. Since x2+1≥1x^2 + 1 \ge 1, 0<2x2+1≤20 < \frac{2}{x^2 + 1} \le 2, with 22 reached only at x=0x = 0. So −1≤q(x)<1-1 \le q(x) < 1: the absolute minimum is q(0)=−1q(0) = -1. The values approach 11 as ∣x∣→∞|x| \to \infty but never reach it, since 2x2+1>0\frac{2}{x^2 + 1} > 0: NO absolute maximum. The interval is not a finite closed interval, so the EVT was never going to help: the rewriting did the work. Check with Theorem 2: q′(x)=4x(x2+1)2q'(x) = \frac{4x}{(x^2 + 1)^2} vanishes only at 00, the minimum.

Exercise 8: Five statements to correct

Each statement below was written by a student on a MATH 203 assignment, and each is false. Say what is wrong, give a counterexample, and write the correct statement, with the definitions of Thomas.

  • a) An endpoint of the interval can never be a local extremum.
  • b) If f′(c)f'(c) does not exist, then ff has a local maximum or a local minimum at cc.
  • c) To find the absolute maximum of ff on [a,b][a, b], evaluate f′f' at the critical points and at the endpoints, and take the largest value.
  • d) A function continuous on [a,b][a, b] attains its absolute maximum at exactly one point.
  • e) My calculator gives an error for (−8)2/3(-8)^{2/3}, so x=−8x = -8 is not in the domain of f(x)=x2/3f(x) = x^{2/3} and cannot be a candidate.

Type your answers, the page tells you right or wrong 0/6

a)
b)
c)
d)
e)
Show the solution

Answers

  • a) False in Thomas: x\sqrt x on [0,4][0, 4] has a local minimum at the endpoint 00. Endpoints are never CRITICAL points.
  • b) False: x1/3x^{1/3} at 00. An undefined f′f' only makes cc a critical point, a candidate.
  • c) False: x3−3xx^3 - 3x on [0,2][0, 2] gives 99 from f′(2)f'(2); the true maximum is f(2)=2f(2) = 2. Evaluate ff, not f′f'.
  • d) False: sin⁡x\sin x on [0,4π][0, 4\pi] reaches 11 at π2\frac{\pi}{2} and 5π2\frac{5\pi}{2}. The EVT gives existence, not uniqueness.
  • e) False: (−8)2/3=((−8)1/3)2=4(-8)^{2/3} = \left((-8)^{1/3}\right)^2 = 4. The domain of x2/3x^{2/3} is all real numbers.

a) FALSE with the definitions of Thomas. f(x)=xf(x) = \sqrt x on [0,4][0, 4]: f(x)≥0=f(0)f(x) \ge 0 = f(0) for all xx in [0,δ)[0, \delta), so f(0)=0f(0) = 0 is a local minimum value at the endpoint 00, and f(4)=2f(4) = 2 is a local maximum value at the endpoint 44. Thomas compares f(c)f(c) with the values of ff at the points of its DOMAIN near cc, which at an endpoint means a half-open interval. Correct statement: an endpoint can be a local extremum, but it is never a critical point, since critical points are interior points. It is always a candidate for an absolute extremum.

b) FALSE. f(x)=x1/3f(x) = x^{1/3}: f′(x)=13x2/3f'(x) = \frac{1}{3x^{2/3}} is undefined at 00 (vertical tangent), yet x1/3<0x^{1/3} < 0 for x<0x < 0 and x1/3>0x^{1/3} > 0 for x>0x > 0, so f(0)=0f(0) = 0 is neither a local maximum nor a local minimum. Correct statement: if f′(c)f'(c) does not exist and cc is an interior point of the domain, then cc is a CRITICAL POINT, that is, a candidate; whether it carries an extremum must be decided separately. Undefined is no more a verdict than zero.

c) FALSE. Take f(x)=x3−3xf(x) = x^3 - 3x on [0,2][0, 2]: f′(x)=3x2−3f'(x) = 3x^2 - 3, critical point 11. The student's list is f′(0)=−3f'(0) = -3, f′(1)=0f'(1) = 0, f′(2)=9f'(2) = 9, and the answer max 99. But the question is about the VALUES of ff: f(0)=0f(0) = 0, f(1)=−2f(1) = -2, f(2)=8−6=2f(2) = 8 - 6 = 2. Absolute maximum 22 at x=2x = 2, absolute minimum −2-2 at x=1x = 1. At a critical point f′f' is 00 or undefined by definition, so a table of f′f' carries no information. Correct statement: evaluate ff at the critical points and the endpoints, and take the largest value OF ff.

d) FALSE. f(x)=sin⁡xf(x) = \sin x is continuous on [0,4π][0, 4\pi] and reaches its absolute maximum value 11 at x=π2x = \frac{\pi}{2} AND at x=5π2x = \frac{5\pi}{2}; a constant function reaches it at every point. Correct statement (Theorem 1): a function continuous on a closed interval [a,b][a, b] attains an absolute maximum value at LEAST once. The theorem guarantees existence, not uniqueness, which is why a complete answer lists every place where the maximum value occurs.

e) FALSE. x2/3=(x1/3)2x^{2/3} = \left(x^{1/3}\right)^2, and the cube root of a negative number exists: (−8)1/3=−2(-8)^{1/3} = -2, so (−8)2/3=(−2)2=4(-8)^{2/3} = (-2)^2 = 4. Many scientific calculators refuse a fractional power of a negative base, because they compute it with logarithms; the number exists all the same. Correct statement: the domain of x2/3x^{2/3} is all real numbers, and at a negative xx it is computed as (x1/3)2\left(x^{1/3}\right)^2 or (x2)1/3\left(x^2\right)^{1/3}. On an interval such as [−8,1][-8, 1], x=−8x = -8 is an endpoint and must be evaluated: f(−8)=4f(-8) = 4 is in fact the absolute maximum there.

Exercise 9: The temperature over a day: warmest, coldest, and when it warms fastest

A weather station fits the outdoor temperature on a spring day by T(t)=10−1.92t+0.3t2−0.01t3T(t) = 10 - 1.92t + 0.3t^2 - 0.01t^3, in degrees Celsius, where tt is the time in hours after midnight, 0≤t≤200 \le t \le 20 (until 8 p.m.). The figure shows the model.

The decimal coefficients are part of the question: factor them out before solving anything. A calculator is allowed for the values; give temperatures to 2 decimals.

246810121416182024681012141618y = T(t)t (hours after midnight)T (°C)
  • a) Explain why TT has an absolute maximum and an absolute minimum on [0,20][0, 20], then find its critical points.
  • b) Find the highest and lowest temperatures of the day between midnight and 8 p.m., and the clock times at which they occur.
  • c) The rate of change T′(t)T'(t) is in degrees per hour. Find the absolute maximum and minimum of T′T' on [0,20][0, 20]: when does the air warm fastest, and when does it cool fastest?
  • d) A student uses the model on [0,24][0, 24]. Find the absolute minimum of TT on [0,24][0, 24] and comment on the model.
  • e) On [0,20][0, 20], how many local extrema does TT have, endpoints included as in Thomas? Justify the endpoints with the closed interval method on [0,4][0, 4] and on [16,20][16, 20].

Type your answers, the page tells you right or wrong 0/11

a)
b)
c)
d)
e)
Show the solution

Answers

  • a) TT is a polynomial, continuous on [0,20][0, 20]; T′(t)=−0.03(t−4)(t−16)T'(t) = -0.03(t - 4)(t - 16), critical points 44 and 1616
  • b) Max 15.1215.12 °C at t=16t = 16 (4 p.m.); min 6.486.48 °C at t=4t = 4 (4 a.m.)
  • c) Warms fastest at t=10t = 10 (10 a.m.), 1.081.08 °C/h; cools fastest at t=0t = 0 and t=20t = 20, −1.92-1.92 °C/h
  • d) Min −1.52-1.52 °C at t=24t = 24: a drop of 1313 degrees after 8 p.m., the model is not reliable past t=20t = 20
  • e) Four: local max at t=0t = 0 and t=16t = 16, local min at t=4t = 4 and t=20t = 20

a) TT is a polynomial, so it is continuous on the closed interval [0,20][0, 20], and by the Extreme Value Theorem it attains an absolute maximum and an absolute minimum there. T′(t)=−1.92+0.6t−0.03t2T'(t) = -1.92 + 0.6t - 0.03t^2. The algebraic gesture: factor out −0.03-0.03 BEFORE solving, dividing each coefficient: −1.92−0.03=64\frac{-1.92}{-0.03} = 64 and 0.6−0.03=−20\frac{0.6}{-0.03} = -20. So T′(t)=−0.03(t2−20t+64)=−0.03(t−4)(t−16)T'(t) = -0.03\left(t^2 - 20t + 64\right) = -0.03(t - 4)(t - 16), since 4×16=644 \times 16 = 64 and 4+16=204 + 16 = 20. Critical points: t=4t = 4 and t=16t = 16, both in (0,20)(0, 20). Applying the quadratic formula to −0.03t2+0.6t−1.92=0-0.03t^2 + 0.6t - 1.92 = 0 directly gives the same roots, with three times as many chances of a sign slip.

b) Values: T(0)=10T(0) = 10; T(4)=10−7.68+4.8−0.64=6.48T(4) = 10 - 7.68 + 4.8 - 0.64 = 6.48; T(16)=10−30.72+76.8−40.96=15.12T(16) = 10 - 30.72 + 76.8 - 40.96 = 15.12; T(20)=10−38.4+120−80=11.6T(20) = 10 - 38.4 + 120 - 80 = 11.6. The highest temperature is 15.1215.12 °C, at t=16t = 16, that is 4 p.m.; the lowest is 6.486.48 °C, at t=4t = 4, that is 4 a.m. The answer to where is a clock time here, and a value of tt converted to it: t=16t = 16 is not 4 o'clock without p.m.

c) T′T' is a polynomial, continuous on [0,20][0, 20]: apply the same method to T′T'. Its derivative is T′′(t)=0.6−0.06tT''(t) = 0.6 - 0.06t, zero at t=10t = 10. Values of T′T': T′(0)=−1.92T'(0) = -1.92; T′(10)=−0.03(6)(−6)=1.08T'(10) = -0.03(6)(-6) = 1.08; T′(20)=−0.03(16)(4)=−1.92T'(20) = -0.03(16)(4) = -1.92. The largest rate is 1.081.08 °C per hour, at 10 a.m.: the air warms fastest in mid-morning. The smallest is −1.92-1.92 °C per hour, at t=0t = 0 and t=20t = 20: the fastest cooling happens at both ends of the period. Here the closed interval method is applied to a DIFFERENT function, T′T', whose candidates are the zeros of T′′T'' and the endpoints; nothing about the concavity of TT is needed.

d) On [0,24][0, 24] the critical points are still 44 and 1616, and the new endpoint gives T(24)=10−46.08+172.8−138.24=−1.52T(24) = 10 - 46.08 + 172.8 - 138.24 = -1.52. Candidates: T(0)=10T(0) = 10, T(4)=6.48T(4) = 6.48, T(16)=15.12T(16) = 15.12, T(24)=−1.52T(24) = -1.52. The absolute minimum becomes −1.52-1.52 °C at midnight at the end of the day, while T(4)=6.48T(4) = 6.48 stays a local minimum. Comment: the model predicts a drop of about 1313 degrees in four hours after 8 p.m., and a cubic with a negative leading coefficient eventually plunges without bound. The fit is valid on [0,20][0, 20] only: an absolute extremum found at an endpoint where the model is extrapolated describes the formula, not the weather.

e) On [0,4][0, 4], TT is continuous and has no critical point inside (0,4)(0, 4), so its extrema there are at the endpoints: T(0)=10T(0) = 10 is the maximum on [0,4][0, 4]. Hence T(t)≤T(0)T(t) \le T(0) on [0,4)[0, 4), a half-open interval of the domain containing 00: a local maximum at the endpoint 00, in Thomas's sense. Similarly on [16,20][16, 20], no critical point inside, T(16)=15.12T(16) = 15.12 and T(20)=11.6T(20) = 11.6: T(20)T(20) is the minimum on [16,20][16, 20], a local minimum at the endpoint 2020. Inside, T(4)=6.48T(4) = 6.48 is the absolute minimum on [0,20][0, 20] at an interior point, hence a local minimum, and T(16)=15.12T(16) = 15.12 the absolute maximum, hence a local maximum. Four local extrema: maxima at t=0t = 0 and t=16t = 16, minima at t=4t = 4 and t=20t = 20.

Exercise 10: A final exam question: a parameter inside the formula, and a cubic to factor

Let a>0a > 0 be a constant and f(x)=x3−3a2xf(x) = x^3 - 3a^2x on the interval [0,2][0, 2].

The shape of a long final exam question: the list of candidates depends on aa, so the answer depends on aa, and every case must be argued. The algebra carries the marks: factor f′f', then factor a cubic in aa.

  • a) Find the critical points of ff in terms of aa. For which values of aa does one of them lie in (0,2)(0, 2)?
  • b) Take a=1a = 1. Find the absolute maximum and minimum of ff on [0,2][0, 2].
  • c) Take a=3a = 3. Find the absolute maximum and minimum of ff on [0,2][0, 2].
  • d) Let 0<a<20 < a < 2. Show that f(2)−f(a)=2(a+1)(a−2)2f(2) - f(a) = 2(a + 1)(a - 2)^2, and deduce the absolute minimum of ff on [0,2][0, 2]. What is the double root of a3−3a2+4a^3 - 3a^2 + 4?
  • e) For which values of aa is the absolute maximum of ff on [0,2][0, 2] attained at x=0x = 0 rather than at x=2x = 2? Give the boundary value to 4 decimals.

Type your answers, the page tells you right or wrong 0/8

a)
Values of aa with a critical point in (0,2)(0, 2) ,
b)
c)
d)
e)
Show the solution

Answers

  • a) x=ax = a and x=−ax = -a; only x=ax = a can be in (0,2)(0, 2), when 0<a<20 < a < 2
  • b) Max 22 at x=2x = 2; min −2-2 at x=1x = 1
  • c) Max 00 at x=0x = 0; min −46-46 at x=2x = 2
  • d) f(2)−f(a)=2(a3−3a2+4)=2(a+1)(a−2)2≥0f(2) - f(a) = 2(a^3 - 3a^2 + 4) = 2(a + 1)(a - 2)^2 \ge 0: min −2a3-2a^3 at x=ax = a; double root a=2a = 2
  • e) a≥23≈1.1547a \ge \frac{2}{\sqrt 3} \approx 1.1547 (tie at x=0x = 0 and x=2x = 2 when a=23a = \frac{2}{\sqrt 3})

a) ff is a polynomial, continuous on [0,2][0, 2] for every aa. f′(x)=3x2−3a2=3(x−a)(x+a)f'(x) = 3x^2 - 3a^2 = 3(x - a)(x + a): critical points x=ax = a and x=−ax = -a. Since a>0a > 0, −a<0-a < 0 is never in (0,2)(0, 2): rejected in every case. And x=ax = a lies in (0,2)(0, 2) exactly when 0<a<20 < a < 2. Treat aa as a number while differentiating: the derivative of 3a2x3a^2x with respect to xx is 3a23a^2, not 6ax6ax.

b) a=1a = 1: f(x)=x3−3xf(x) = x^3 - 3x, candidates 00, 11, 22. f(0)=0f(0) = 0, f(1)=−2f(1) = -2, f(2)=8−6=2f(2) = 8 - 6 = 2. Absolute maximum 22 at x=2x = 2, absolute minimum −2-2 at x=1x = 1.

c) a=3a = 3: the critical point 33 is outside (0,2)(0, 2), rejected, and so is −3-3. Only the endpoints remain: f(0)=0f(0) = 0, f(2)=8−3⋅9⋅2=8−54=−46f(2) = 8 - 3 \cdot 9 \cdot 2 = 8 - 54 = -46. Absolute maximum 00 at x=0x = 0, absolute minimum −46-46 at x=2x = 2. A student who evaluates at x=3x = 3 gets f(3)=27−81=−54f(3) = 27 - 81 = -54, a value ff never takes on [0,2][0, 2].

d) For 0<a<20 < a < 2 the candidates are 00, aa, 22, with f(0)=0f(0) = 0, f(a)=a3−3a3=−2a3f(a) = a^3 - 3a^3 = -2a^3 and f(2)=8−6a2f(2) = 8 - 6a^2. Then f(2)−f(a)=8−6a2+2a3=2(a3−3a2+4)f(2) - f(a) = 8 - 6a^2 + 2a^3 = 2\left(a^3 - 3a^2 + 4\right). To factor the cubic, look for an integer root among the divisors of 44: a=−1a = -1 gives −1−3+4=0-1 - 3 + 4 = 0. Division by a+1a + 1 gives a3−3a2+4=(a+1)(a2−4a+4)=(a+1)(a−2)2a^3 - 3a^2 + 4 = (a + 1)\left(a^2 - 4a + 4\right) = (a + 1)(a - 2)^2; its double root is a=2a = 2. For a>0a > 0 both factors are ≥0\ge 0, so f(2)≥f(a)f(2) \ge f(a); and f(0)=0>−2a3=f(a)f(0) = 0 > -2a^3 = f(a). The absolute minimum is −2a3-2a^3, at x=ax = a. For a=1a = 1 this gives −2-2, as in b).

e) For 0<a<20 < a < 2 the maximum is the larger of f(0)=0f(0) = 0 and f(2)=8−6a2f(2) = 8 - 6a^2 (the interior candidate gives the minimum by d)). f(0)≥f(2)f(0) \ge f(2) means 6a2≥86a^2 \ge 8, a2≥43a^2 \ge \frac{4}{3}, a≥23=233≈1.1547a \ge \frac{2}{\sqrt 3} = \frac{2\sqrt 3}{3} \approx 1.1547. For a≥2a \ge 2, by c), the only candidates are the endpoints and f(2)=8−6a2<0f(2) = 8 - 6a^2 < 0, so the maximum is again at x=0x = 0. Conclusion: the absolute maximum 00 is at x=0x = 0 when a≥23a \ge \frac{2}{\sqrt 3} (at both endpoints in the tie a=23a = \frac{2}{\sqrt 3}), and it is 8−6a28 - 6a^2 at x=2x = 2 when 0<a<230 < a < \frac{2}{\sqrt 3}. The figure of the solution shows the two regimes, a=1a = 1 and a=1.5a = 1.5.

0.511.522.5-8-7-6-5-4-3-2-1123a = 1a = 1.5

© Ahmed Squalli Houssaini. Exercise set published at www.letuteurscientifique.ca/en/exercices/math203-extreme-values. Free for personal and classroom use; republishing it elsewhere requires written permission (legal notice).

See also

Struggling with MATH 203?

I tutor first-year calculus at Concordia and McGill, in English or in French, in Montreal or online. Get in touch for a first session.

Site by Studio Squalli