MATH 203 Calculus I • Concordia University, Montreal

Corrected exercises: inverse functions and logarithms (MATH 203)

This is the corrected exercise set for inverse functions and logarithms in MATH 203, Differential and Integral Calculus I, at Concordia University: section 1.6 of Thomas' Calculus, without the inverse trigonometric functions, which come later in the course. Everything in the chapters that follow leans on it: the derivative of ln⁡x\ln x, logarithmic differentiation and every growth model assume that the laws of logarithms and the domain of ln⁡\ln are automatic. A scientific calculator is allowed, as on the exams, so exact answers come first and decimals follow, rounding announced.

The thread running through the set is an algebra gesture, because that is where MATH 203 marks are lost, not in the idea of an inverse. When the unknown appears twice, xx in a numerator and a denominator, exe^x and e−xe^{-x}, xln⁡3x\ln 3 and xln⁡7x\ln 7, it must be COLLECTED on one side and FACTORED out before anything else. And an inverse swaps inputs and outputs, so every answer ends with a domain check: the piece where ff is one-to-one, the positive arguments of a logarithm, the range that becomes a domain.

The traps named in the solutions: xx left on both sides of an inverse formula, a ±\pm left in an inverse, a restriction forgotten, f−1(5)f^{-1}(5) confused with 1f(5)\frac{1}{f(5)}, ln⁡(a+b)=ln⁡a+ln⁡b\ln(a + b) = \ln a + \ln b, ln⁡aln⁡b=ln⁡ab\frac{\ln a}{\ln b} = \ln\frac{a}{b}, (ln⁡x)2=2ln⁡x(\ln x)^2 = 2\ln x, a change of base upside down, e2ln⁡x=x2e^{2\ln x} = x^2 for negative xx, a candidate kept that makes an argument negative, a valid negative root rejected by reflex, the graph reflected in the wrong axis, a half-life treated linearly, and decibels added.

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

  • • One-to-one: f(a)=f(b)⇒a=bf(a) = f(b) \Rightarrow a = b (horizontal line test). Only a one-to-one function has an inverse: f−1(b)=a  ⟺  f(a)=bf^{-1}(b) = a \iff f(a) = b.
  • • Domain of f−1f^{-1} = range of ff; range of f−1f^{-1} = domain of ff. The graph of f−1f^{-1} is the reflection of the graph of ff in the line y=xy = x.
  • • To find f−1f^{-1}: write y=f(x)y = f(x), clear denominators, COLLECT the xx terms, factor xx out, divide, then rename.
  • • log⁡ax=y  ⟺  ay=x\log_a x = y \iff a^y = x, for x>0x > 0 only. log⁡a(xy)=log⁡ax+log⁡ay\log_a(xy) = \log_a x + \log_a y, log⁡axy=log⁡ax−log⁡ay\log_a\frac{x}{y} = \log_a x - \log_a y, log⁡axr=rlog⁡ax\log_a x^r = r\log_a x, for x,y>0x, y > 0.
  • • Change of base: log⁡ax=ln⁡xln⁡a\log_a x = \frac{\ln x}{\ln a}. No law for ln⁡(x+y)\ln(x + y), (ln⁡x)(ln⁡y)(\ln x)(\ln y) or ln⁡xln⁡y\frac{\ln x}{\ln y}.
  • • ln⁡(ex)=x\ln(e^x) = x for every xx; eln⁡x=xe^{\ln x} = x for x>0x > 0 only. Every candidate of a log equation is tested in the ORIGINAL equation.

Part A: the basics (/50)

Exercise 1: One-to-one functions: the horizontal line test and the algebraic test

A function ff is one-to-one on a domain DD when f(a)≠f(b)f(a) \ne f(b) whenever a≠ba \ne b in DD: it never takes the same value twice. On a graph, this is the horizontal line test: no horizontal line meets the curve more than once. Only a one-to-one function has an inverse, and f−1(b)=af^{-1}(b) = a means exactly f(a)=bf(a) = b.

The figure shows the graph of p(x)=x3−3xp(x) = x^3 - 3x, the horizontal line y=2y = 2, and three marked points. The curve turns at (−1,2)(-1, 2) and at (1,−2)(1, -2).

-3-2-1123-4-3-2-11234(-1, 2)(2, 2)(1, -2)y = 2y = p(x)
  • a) Using the figure, find two inputs with the same output for pp and check them by computing. Is pp one-to-one on R\mathbb{R}? How many solutions does the equation p(x)=1p(x) = 1 have?
  • b) Give the largest interval containing 33 on which pp is one-to-one, and the largest interval containing 00. Then prove the first answer by algebra: for 1≤a<b1 \le a < b, factor p(b)−p(a)p(b) - p(a) and find its sign.
  • c) Which of these functions are one-to-one on their domain? Prove it, or give two inputs with the same output: q1(x)=1+e−xq_1(x) = 1 + e^{-x}, q2(x)=x4−1q_2(x) = x^4 - 1, q3(x)=x∣x∣q_3(x) = x|x|, q4(x)=16−x2q_4(x) = \sqrt{16 - x^2}.
  • d) A one-to-one function ff is known only through its table: f(−2)=9f(-2) = 9, f(0)=5f(0) = 5, f(1)=4f(1) = 4, f(3)=−1f(3) = -1, f(6)=−7f(6) = -7. Find f−1(4)f^{-1}(4), f−1(−7)f^{-1}(-7), f−1(f(3))f^{-1}(f(3)), f−1(5)f^{-1}(5) and 1f(0)\frac{1}{f(0)}.
  • e) True or false? (i) A function that is increasing on its whole domain is one-to-one. (ii) A one-to-one function is increasing on its whole domain or decreasing on its whole domain.

Type your answers, the page tells you right or wrong 0/17

a)
b)
Largest interval containing 33 ,
Largest interval containing 00 ,
c)
d)
e)
Show the solution

Answers

  • a) p(−1)=p(2)=2p(-1) = p(2) = 2, so pp is not one-to-one; p(x)=1p(x) = 1 has 33 solutions
  • b) [1,∞)[1, \infty) and [−1,1][-1, 1]; p(b)−p(a)=(b−a)(a2+ab+b2−3)>0p(b) - p(a) = (b - a)(a^2 + ab + b^2 - 3) > 0
  • c) q1q_1 and q3q_3 are one-to-one; q2(−1)=q2(1)=0q_2(-1) = q_2(1) = 0 and q4(−4)=q4(4)=0q_4(-4) = q_4(4) = 0
  • d) f−1(4)=1f^{-1}(4) = 1, f−1(−7)=6f^{-1}(-7) = 6, f−1(f(3))=3f^{-1}(f(3)) = 3, f−1(5)=0f^{-1}(5) = 0, 1f(0)=15\frac{1}{f(0)} = \frac{1}{5}
  • e) (i) true; (ii) false: 1x\frac{1}{x} is one-to-one but not decreasing on its whole domain

a) The line y=2y = 2 meets the curve at the two marked points (−1,2)(-1, 2) and (2,2)(2, 2). Check: p(−1)=−1+3=2p(-1) = -1 + 3 = 2 and p(2)=8−6=2p(2) = 8 - 6 = 2. Two different inputs give the same output, so pp is NOT one-to-one on R\mathbb{R}, and it has no inverse there. For p(x)=1p(x) = 1, draw the line y=1y = 1 in your head: it lies between the low turning value −2-2 and the high one 22, so it cuts the three pieces of the curve, and the equation has 33 solutions. The values p(−2)=−2p(-2) = -2, p(−1)=2p(-1) = 2, p(1)=−2p(1) = -2, p(2)=2p(2) = 2 alternate around 11, which confirms one crossing in each of the three intervals. One counterexample settles a no; the figure is enough to find it, the computation is what proves it.

b) To the right of the turning point (1,−2)(1, -2) the curve only rises, so [1,∞)[1, \infty) is the largest interval containing 33 on which pp is one-to-one; between the two turning points it only falls, so for 00 the answer is [−1,1][-1, 1], where pp takes each value of [−2,2][-2, 2] once. The algebraic proof uses the gesture of the whole chapter, FACTOR before judging a sign: p(b)−p(a)=(b3−a3)−3(b−a)=(b−a)(a2+ab+b2)−3(b−a)=(b−a)(a2+ab+b2−3)p(b) - p(a) = (b^3 - a^3) - 3(b - a) = (b - a)(a^2 + ab + b^2) - 3(b - a) = (b - a)(a^2 + ab + b^2 - 3). For 1≤a<b1 \le a < b: a2≥1a^2 \ge 1, ab>a2≥1ab > a^2 \ge 1 and b2>1b^2 > 1, so the second factor is positive, and so is b−ab - a. Hence p(b)>p(a)p(b) > p(a): pp is increasing on [1,∞)[1, \infty), and an increasing function never repeats a value. Without the common factor b−ab - a pulled out, the sign of b3−a3−3b+3ab^3 - a^3 - 3b + 3a cannot be read, and this is where the marks go.

c) q1q_1: if 1+e−a=1+e−b1 + e^{-a} = 1 + e^{-b} then e−a=e−be^{-a} = e^{-b}, and since the exponential is one-to-one, −a=−b-a = -b, so a=ba = b: one-to-one. q2(−1)=q2(1)=0q_2(-1) = q_2(1) = 0: not one-to-one, an even function never is on a domain symmetric about 00. q3(x)=x∣x∣q_3(x) = x|x| equals x2x^2 for x≥0x \ge 0 and −x2-x^2 for x<0x < 0, so its sign is the sign of xx. If a∣a∣=b∣b∣a|a| = b|b|, then aa and bb have the same sign and a2=∣a∣ ∣a∣=∣b∣ ∣b∣=b2a^2 = |a|\,|a| = |b|\,|b| = b^2, so ∣a∣=∣b∣|a| = |b|, and with the same sign, a=ba = b: one-to-one, although it is built from x2x^2. q4(−4)=q4(4)=0q_4(-4) = q_4(4) = 0: the upper half of a circle of radius 44 fails the horizontal line test.

d) Read the table backwards: f−1(4)f^{-1}(4) is the input whose output is 44, so f−1(4)=1f^{-1}(4) = 1; likewise f−1(−7)=6f^{-1}(-7) = 6 and f−1(5)=0f^{-1}(5) = 0. Next, f−1(f(3))=f−1(−1)=3f^{-1}(f(3)) = f^{-1}(-1) = 3, as the cancellation law f−1(f(x))=xf^{-1}(f(x)) = x predicts on the domain of ff. Finally 1f(0)=15\frac{1}{f(0)} = \frac{1}{5}, which has nothing to do with f−1(5)=0f^{-1}(5) = 0: the −1-1 in f−1f^{-1} means the inverse for composition, never the reciprocal. The reciprocal of the value is written (f(0))−1(f(0))^{-1}, with the brackets.

e) (i) TRUE: if ff is increasing and a<ba < b, then f(a)<f(b)f(a) < f(b), so two different inputs never share an output. (ii) FALSE: g(x)=1xg(x) = \frac{1}{x} is one-to-one, since 1a=1b\frac{1}{a} = \frac{1}{b} gives a=ba = b, but it is not decreasing on its whole domain: −1<1-1 < 1 and g(−1)=−1<1=g(1)g(-1) = -1 < 1 = g(1). It decreases on each of its two pieces, which is not the same thing. The one-to-one property is a statement about VALUES; a monotonicity argument is only one way to prove it.

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: Finding an inverse: collect the unknown, factor it out, then read the domain

Method: write y=f(x)y = f(x), solve for xx, then rename xx and yy. The domain of f−1f^{-1} is the range of ff, and its range is the domain of ff. The algebra always hits the same wall: after clearing a denominator, xx appears in TWO terms. Collect every xx term on one side, factor xx out, then divide. That step, not the idea of an inverse, is where most of the marks are lost.

Check every inverse on one value: compute f(a)=bf(a) = b for an easy aa, then f−1(b)f^{-1}(b) must give back aa.

  • a) Find the inverse of f(x)=1−3xx+4f(x) = \frac{1 - 3x}{x + 4}, with its domain and range. Compute f−1(1)f^{-1}(1) and check it in ff.
  • b) Find the inverse of g(x)=2−x−1g(x) = 2 - \sqrt{x - 1}, x≥1x \ge 1, with its domain. Compute g−1(0)g^{-1}(0) and g−1(−1)g^{-1}(-1).
  • c) Find the inverse of h(x)=1+2ln⁡(x+3)h(x) = 1 + 2\ln(x + 3), with its domain and range. Compute h−1(1)h^{-1}(1) and h−1(5)h^{-1}(5).
  • d) Find the inverse of m(x)=2x+12x−1m(x) = \frac{2^x + 1}{2^x - 1}, x≠0x \ne 0, and its domain. Compute m−1(3)m^{-1}(3) and m−1(−3)m^{-1}(-3), and say whether m−1(0)m^{-1}(0) exists.
  • e) For the function of a), simplify the complex fraction f(f−1(x))f(f^{-1}(x)) and show that it equals xx. For which xx is this valid?

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

a)
b)
Domain of g−1g^{-1} ,
c)
Range of h−1h^{-1} ,
d)
e)
Show the solution

Answers

  • a) f−1(x)=1−4xx+3f^{-1}(x) = \frac{1 - 4x}{x + 3}, domain x≠−3x \ne -3, range y≠−4y \ne -4; f−1(1)=−34f^{-1}(1) = -\frac{3}{4}
  • b) g−1(x)=1+(2−x)2g^{-1}(x) = 1 + (2 - x)^2 for x≤2x \le 2; g−1(0)=5g^{-1}(0) = 5, g−1(−1)=10g^{-1}(-1) = 10
  • c) h−1(x)=e(x−1)/2−3h^{-1}(x) = e^{(x - 1)/2} - 3, domain R\mathbb{R}, range (−3,∞)(-3, \infty); h−1(1)=−2h^{-1}(1) = -2, h−1(5)=e2−3h^{-1}(5) = e^2 - 3
  • d) m−1(x)=log⁡2x+1x−1m^{-1}(x) = \log_2 \frac{x + 1}{x - 1} for x<−1x < -1 or x>1x > 1; m−1(3)=1m^{-1}(3) = 1, m−1(−3)=−1m^{-1}(-3) = -1; m−1(0)m^{-1}(0) does not exist
  • e) f(f−1(x))=13x13=xf(f^{-1}(x)) = \frac{13x}{13} = x for every x≠−3x \ne -3

a) Write y=1−3xx+4y = \frac{1 - 3x}{x + 4} and clear the denominator: y(x+4)=1−3xy(x + 4) = 1 - 3x, that is xy+4y=1−3xxy + 4y = 1 - 3x. Now the gesture: the unknown xx is in two terms, xyxy and −3x-3x. Collect them on the left, everything else on the right: xy+3x=1−4yxy + 3x = 1 - 4y. Factor: x(y+3)=1−4yx(y + 3) = 1 - 4y, and divide, which requires y≠−3y \ne -3: x=1−4yy+3x = \frac{1 - 4y}{y + 3}. Renaming, f−1(x)=1−4xx+3f^{-1}(x) = \frac{1 - 4x}{x + 3}. Its domain is x≠−3x \ne -3, and that is the range of ff: the value −3-3 is never taken, since 1−3xx+4=−3\frac{1 - 3x}{x + 4} = -3 would give 1−3x=−3x−121 - 3x = -3x - 12, that is 1=−121 = -12. Its range is y≠−4y \ne -4, the domain of ff. Then f−1(1)=−34f^{-1}(1) = \frac{-3}{4}, and the check: f(−34)=1+94134=13/413/4=1f\left(-\frac{3}{4}\right) = \frac{1 + \frac{9}{4}}{\frac{13}{4}} = \frac{13/4}{13/4} = 1. The classic wrong line is x=1−3x−4yyx = \frac{1 - 3x - 4y}{y}: xx is still on both sides, and nothing has been solved.

b) gg is decreasing, so it is one-to-one. Its range: x−1\sqrt{x - 1} takes every value in [0,∞)[0, \infty), so gg takes every value in (−∞,2](-\infty, 2]. Solve y=2−x−1y = 2 - \sqrt{x - 1}: x−1=2−y\sqrt{x - 1} = 2 - y. A square root is never negative, so this step REQUIRES 2−y≥02 - y \ge 0, that is y≤2y \le 2. Squaring, x−1=(2−y)2x - 1 = (2 - y)^2, so g−1(x)=1+(2−x)2g^{-1}(x) = 1 + (2 - x)^2 for x≤2x \le 2. The condition is part of the formula: 1+(2−x)21 + (2 - x)^2 on all of R\mathbb{R} is a full parabola, which fails the horizontal line test and is the inverse of nothing. Then g−1(0)=1+4=5g^{-1}(0) = 1 + 4 = 5, checked by g(5)=2−2=0g(5) = 2 - 2 = 0, and g−1(−1)=1+9=10g^{-1}(-1) = 1 + 9 = 10, checked by g(10)=2−3=−1g(10) = 2 - 3 = -1.

c) The domain of hh is x>−3x > -3. Isolate the logarithm first, then undo it: y=1+2ln⁡(x+3)y = 1 + 2\ln(x + 3) gives ln⁡(x+3)=y−12\ln(x + 3) = \frac{y - 1}{2}, so x+3=e(y−1)/2x + 3 = e^{(y - 1)/2} and x=e(y−1)/2−3x = e^{(y - 1)/2} - 3. So h−1(x)=e(x−1)/2−3h^{-1}(x) = e^{(x - 1)/2} - 3, defined for every real xx (the range of hh is R\mathbb{R}, like that of ln⁡\ln), with range (−3,∞)(-3, \infty), the domain of hh, because e(x−1)/2>0e^{(x-1)/2} > 0. Values: h−1(1)=e0−3=−2h^{-1}(1) = e^0 - 3 = -2, checked by h(−2)=1+2ln⁡1=1h(-2) = 1 + 2\ln 1 = 1; and h−1(5)=e2−3≈4.3891h^{-1}(5) = e^2 - 3 \approx 4.3891. The frequent slip is to exponentiate before isolating: ey=e⋅2(x+3)e^y = e \cdot 2(x + 3) confuses e2ln⁡(x+3)=(x+3)2e^{2\ln(x+3)} = (x + 3)^2 with a product.

d) Write y=2x+12x−1y = \frac{2^x + 1}{2^x - 1} and clear the denominator: y⋅2x−y=2x+1y \cdot 2^x - y = 2^x + 1. The unknown quantity 2x2^x appears in two terms: collect them, y⋅2x−2x=y+1y \cdot 2^x - 2^x = y + 1, factor, 2x(y−1)=y+12^x(y - 1) = y + 1, and divide (y≠1y \ne 1): 2x=y+1y−12^x = \frac{y + 1}{y - 1}. Now the domain appears by itself: 2x>02^x > 0, so we need y+1y−1>0\frac{y + 1}{y - 1} > 0, that is y<−1y < -1 or y>1y > 1 (sign table of y+1y + 1 and y−1y - 1). Then x=log⁡2y+1y−1x = \log_2 \frac{y + 1}{y - 1} and m−1(x)=log⁡2x+1x−1m^{-1}(x) = \log_2 \frac{x + 1}{x - 1} on (−∞,−1)∪(1,∞)(-\infty, -1) \cup (1, \infty), which is therefore the range of mm. Values: m−1(3)=log⁡22=1m^{-1}(3) = \log_2 2 = 1, checked by m(1)=31=3m(1) = \frac{3}{1} = 3; m−1(−3)=log⁡2−2−4=log⁡212=−1m^{-1}(-3) = \log_2 \frac{-2}{-4} = \log_2 \frac{1}{2} = -1, checked by m(−1)=3/2−1/2=−3m(-1) = \frac{3/2}{-1/2} = -3. And m−1(0)m^{-1}(0) does NOT exist: it would be log⁡2(−1)\log_2(-1), and indeed m(x)=0m(x) = 0 would need 2x=−12^x = -1.

e) Substitute f−1(x)=1−4xx+3f^{-1}(x) = \frac{1 - 4x}{x + 3} into ff: f(f−1(x))=1−3⋅1−4xx+31−4xx+3+4f(f^{-1}(x)) = \dfrac{1 - 3\cdot\frac{1 - 4x}{x + 3}}{\frac{1 - 4x}{x + 3} + 4}. The algebra gesture for a complex fraction: multiply the top AND the bottom by the inner denominator x+3x + 3. Top: (x+3)−3(1−4x)=x+3−3+12x=13x(x + 3) - 3(1 - 4x) = x + 3 - 3 + 12x = 13x. Bottom: (1−4x)+4(x+3)=1−4x+4x+12=13(1 - 4x) + 4(x + 3) = 1 - 4x + 4x + 12 = 13. So f(f−1(x))=13x13=xf(f^{-1}(x)) = \frac{13x}{13} = x. It is valid wherever f−1(x)f^{-1}(x) exists, that is for every x≠−3x \ne -3. The two usual errors are multiplying only the top by x+3x + 3, and distributing the −3-3 to the first term only, x+3−3−4xx + 3 - 3 - 4x; both break the cancellation that proves the inverse is right.

Exercise 3: The graph of the inverse: swap, reflect, and restrict before inverting

The point (a,b)(a, b) is on the graph of ff exactly when (b,a)(b, a) is on the graph of f−1f^{-1}. So the graph of f−1f^{-1} is the reflection of the graph of ff in the line y=xy = x, provided both axes carry the same scale. A function that is not one-to-one can still be inverted on a piece of its domain where it is: the restriction is then part of the answer.

The figure shows the graph of f(x)=2x−3f(x) = 2^x - 3, three points on it, and the horizontal line y=−3y = -3, which the graph approaches on the left without ever reaching it.

-4-3-2-1123-4-3-2-112(0, -2)(1, -1)(2, 1)y = -3y = f(x)
  • a) Using the figure, find f−1(−1)f^{-1}(-1) and f−1(1)f^{-1}(1), and solve f(x)=−2f(x) = -2.
  • b) Give the domain and range of ff and of f−1f^{-1}. Which line does the graph of f−1f^{-1} approach, and why?
  • c) Find a formula for f−1f^{-1}. Compute f−1(4)f^{-1}(4) exactly, then to four decimals with your calculator, which has only the ln⁡\ln and log⁡\log keys.
  • d) The function r(x)=e1−x2r(x) = e^{1 - x^2} is not one-to-one. Restrict it to (−∞,0](-\infty, 0], find the inverse of the restriction and its domain, then compute r−1(1)r^{-1}(1) and r−1(e−3)r^{-1}(e^{-3}).
  • e) Sketch f−1f^{-1} by reflecting the figure, and give the points where its graph crosses the two axes.

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

a)
b)
Domain of f−1f^{-1} ,
Range of f−1f^{-1} ,
c)
d)
e)
Show the solution

Answers

  • a) f−1(−1)=1f^{-1}(-1) = 1, f−1(1)=2f^{-1}(1) = 2; f(x)=−2f(x) = -2 gives x=0x = 0
  • b) ff: domain R\mathbb{R}, range (−3,∞)(-3, \infty); f−1f^{-1}: domain (−3,∞)(-3, \infty), range R\mathbb{R}; f−1f^{-1} approaches the vertical line x=−3x = -3
  • c) f−1(x)=log⁡2(x+3)f^{-1}(x) = \log_2(x + 3); f−1(4)=log⁡27=ln⁡7ln⁡2≈2.8074f^{-1}(4) = \log_2 7 = \frac{\ln 7}{\ln 2} \approx 2.8074
  • d) r−1(x)=−1−ln⁡xr^{-1}(x) = -\sqrt{1 - \ln x} on (0,e](0, e]; r−1(1)=−1r^{-1}(1) = -1, r−1(e−3)=−2r^{-1}(e^{-3}) = -2
  • e) xx-intercept (−2,0)(-2, 0), yy-intercept (0,log⁡23)≈(0,1.5850)(0, \log_2 3) \approx (0, 1.5850)

a) f−1(−1)f^{-1}(-1) is the INPUT whose output is −1-1: the marked point (1,−1)(1, -1) gives f−1(−1)=1f^{-1}(-1) = 1. Likewise (2,1)(2, 1) gives f−1(1)=2f^{-1}(1) = 2. Solving f(x)=−2f(x) = -2 is the same question once more, x=f−1(−2)x = f^{-1}(-2), and the point (0,−2)(0, -2) gives x=0x = 0; check, 20−3=−22^0 - 3 = -2. Reading an inverse on a graph means entering on the VERTICAL axis and leaving on the horizontal one: students who enter on the horizontal axis read f(1)=−1f(1) = -1 and f(−1)=−52f(-1) = -\frac{5}{2} instead.

b) ff is defined for every real xx. Since 2x2^x takes every value in (0,∞)(0, \infty), 2x−32^x - 3 takes every value in (−3,∞)(-3, \infty): that is the range of ff. Swap them: f−1f^{-1} has domain (−3,∞)(-3, \infty) and range R\mathbb{R}. The horizontal line y=−3y = -3 that ff approaches on the left is made of points (x,−3)(x, -3); reflected in y=xy = x they become (−3,x)(-3, x), the VERTICAL line x=−3x = -3, which the graph of f−1f^{-1} approaches from the right as it falls. Domain and range swap, and so do horizontal and vertical features.

c) Solve y=2x−3y = 2^x - 3: 2x=y+32^x = y + 3, and apply log⁡2\log_2, legal since y+3>0y + 3 > 0 on the domain: x=log⁡2(y+3)x = \log_2(y + 3). So f−1(x)=log⁡2(x+3)f^{-1}(x) = \log_2(x + 3). Then f−1(4)=log⁡27f^{-1}(4) = \log_2 7, the exact answer. The calculator has no log⁡2\log_2 key, and this is exactly what the change of base formula is for: log⁡27=ln⁡7ln⁡2≈1.9459100.693147≈2.8074\log_2 7 = \frac{\ln 7}{\ln 2} \approx \frac{1.945910}{0.693147} \approx 2.8074. Sanity check: 4<7<84 < 7 < 8, so log⁡27\log_2 7 must lie between 22 and 33; and 22.8074−3≈42^{2.8074} - 3 \approx 4. Typing ln⁡2ln⁡7≈0.3562\frac{\ln 2}{\ln 7} \approx 0.3562 is the upside-down version, and the bracket catches it at once.

d) r(−1)=r(1)=e0=1r(-1) = r(1) = e^0 = 1, so rr is not one-to-one on R\mathbb{R}. On (−∞,0](-\infty, 0], 1−x21 - x^2 increases up to 11, so rr increases from values close to 00 up to r(0)=er(0) = e: its range there is (0,e](0, e], the domain of the inverse. Solve y=e1−x2y = e^{1 - x^2}: take ln⁡\ln, ln⁡y=1−x2\ln y = 1 - x^2, so x2=1−ln⁡yx^2 = 1 - \ln y and x=±1−ln⁡yx = \pm\sqrt{1 - \ln y}. The RESTRICTION chooses the sign: x≤0x \le 0, so x=−1−ln⁡yx = -\sqrt{1 - \ln y}, and r−1(x)=−1−ln⁡xr^{-1}(x) = -\sqrt{1 - \ln x} for 0<x≤e0 < x \le e (the square root needs ln⁡x≤1\ln x \le 1, the same condition). Values: r−1(1)=−1−0=−1r^{-1}(1) = -\sqrt{1 - 0} = -1 and r−1(e−3)=−1+3=−2r^{-1}(e^{-3}) = -\sqrt{1 + 3} = -2; check, r(−2)=e1−4=e−3r(-2) = e^{1 - 4} = e^{-3}. Answering +1+1 gives a number that rr sends to 11 too, but it is not in the restricted domain, so it is not the value of THIS inverse.

e) Each point (a,b)(a, b) of ff becomes (b,a)(b, a): (0,−2)(0, -2), (1,−1)(1, -1), (2,1)(2, 1) become (−2,0)(-2, 0), (−1,1)(-1, 1), (1,2)(1, 2), drawn in the solution figure on equal scales. The graph of f−1f^{-1} crosses the xx-axis at (−2,0)(-2, 0), the image of the yy-intercept (0,−2)(0, -2) of ff; it crosses the yy-axis at (0,log⁡23)(0, \log_2 3), the image of the xx-intercept of ff, where 2x=32^x = 3. So log⁡23=ln⁡3ln⁡2≈1.5850\log_2 3 = \frac{\ln 3}{\ln 2} \approx 1.5850. Intercepts swap roles, exactly like domain and range.

-4-3-2-11234-4-3-2-11234y = f(x)y = f⁻¹(x)y = x

Exercise 4: Laws of logarithms: exact values, condensing, and the laws that do not exist

log⁡ax=y\log_a x = y means ay=xa^y = x (a>0a > 0, a≠1a \ne 1, x>0x > 0), and ln⁡x=log⁡ex\ln x = \log_e x. For x>0x > 0 and y>0y > 0: log⁡a(xy)=log⁡ax+log⁡ay\log_a(xy) = \log_a x + \log_a y, log⁡axy=log⁡ax−log⁡ay\log_a \frac{x}{y} = \log_a x - \log_a y, log⁡axr=rlog⁡ax\log_a x^r = r\log_a x. Change of base: log⁡ax=ln⁡xln⁡a\log_a x = \frac{\ln x}{\ln a}.

There is no law for log⁡a(x+y)\log_a(x + y), none for a product of two logarithms, and none for a quotient of two logarithms. Your scientific calculator has an ln⁡\ln key and a log⁡\log key (base 1010) only.

  • a) Find exactly: log⁡1/381\log_{1/3} 81, log⁡814\log_8 \frac{1}{4}, ln⁡(e2e3)\ln\left(e^2 \sqrt[3]{e}\right), e−2ln⁡5e^{-2\ln 5}, log⁡64+log⁡69\log_6 4 + \log_6 9.
  • b) Write 2ln⁡(x+1)−ln⁡(x2−1)+ln⁡32\ln(x + 1) - \ln(x^2 - 1) + \ln 3 as a single logarithm, simplified, and say for which xx your equality holds. Evaluate the expression at x=3x = 3.
  • c) Expand log⁡28x3y\log_2 \frac{8x^3}{\sqrt y} for x>0x > 0, y>0y > 0 as a sum of simple terms.
  • d) True or false for all a,b>0a, b > 0 (and a,b≠1a, b \ne 1 in (iv))? Give numbers to refute each false one. (i) ln⁡(a+b)=ln⁡a+ln⁡b\ln(a + b) = \ln a + \ln b; (ii) ln⁡(ab)=(ln⁡a)(ln⁡b)\ln(ab) = (\ln a)(\ln b); (iii) ln⁡ab=ln⁡aln⁡b\ln\frac{a}{b} = \frac{\ln a}{\ln b}; (iv) log⁡ab⋅log⁡ba=1\log_a b \cdot \log_b a = 1.
  • e) With your calculator, find log⁡320\log_3 20 and log⁡0.512\log_{0.5} 12 to four decimals. Between which two consecutive integers must log⁡320\log_3 20 lie, before any calculation?

Type your answers, the page tells you right or wrong 0/17

a)
b)
The equality holds for xx in ,
c)
d)
e)
Show the solution

Answers

  • a) −4-4, −23-\frac{2}{3}, 73\frac{7}{3}, 125\frac{1}{25}, 22
  • b) ln⁡3(x+1)x−1\ln\frac{3(x + 1)}{x - 1}, valid for x>1x > 1; at x=3x = 3: ln⁡6≈1.7918\ln 6 \approx 1.7918
  • c) 3+3log⁡2x−12log⁡2y3 + 3\log_2 x - \frac{1}{2}\log_2 y
  • d) (i), (ii), (iii) false (a=b=1a = b = 1; a=b=ea = b = e; a=8a = 8, b=2b = 2); (iv) true
  • e) log⁡320≈2.7268\log_3 20 \approx 2.7268, log⁡0.512≈−3.5850\log_{0.5} 12 \approx -3.5850; 2<log⁡320<32 < \log_3 20 < 3

a) Each logarithm answers a question about an exponent. (13)y=81=34\left(\frac{1}{3}\right)^y = 81 = 3^4 means 3−y=343^{-y} = 3^4, so log⁡1/381=−4\log_{1/3} 81 = -4. 8y=148^y = \frac{1}{4} means 23y=2−22^{3y} = 2^{-2}, so y=−23y = -\frac{2}{3}. ln⁡(e2e1/3)=ln⁡e7/3=73\ln\left(e^2 e^{1/3}\right) = \ln e^{7/3} = \frac{7}{3}: the exponents add BEFORE the logarithm is taken. e−2ln⁡5=eln⁡5−2=5−2=125e^{-2\ln 5} = e^{\ln 5^{-2}} = 5^{-2} = \frac{1}{25}, by the power law then eln⁡u=ue^{\ln u} = u. log⁡64+log⁡69=log⁡636=2\log_6 4 + \log_6 9 = \log_6 36 = 2. Rewriting every base as a power of the same prime is the gesture that makes the exponent visible; the MATH 203 error is to divide the numbers, log⁡814=132\log_8 \frac{1}{4} = \frac{1}{32}.

b) Bring every term inside one logarithm: 2ln⁡(x+1)=ln⁡(x+1)22\ln(x + 1) = \ln(x + 1)^2, so the expression is ln⁡3(x+1)2x2−1\ln\frac{3(x + 1)^2}{x^2 - 1}. Now the algebra step that makes it a simplified answer: FACTOR the difference of squares, x2−1=(x−1)(x+1)x^2 - 1 = (x - 1)(x + 1), and cancel one factor x+1x + 1: ln⁡3(x+1)x−1\ln\frac{3(x + 1)}{x - 1}. Validity: the original needs x+1>0x + 1 > 0 and x2−1>0x^2 - 1 > 0, that is x>1x > 1. The simplified logarithm is also defined for x<−1x < -1, where the original is not, so the equality holds for x>1x > 1 only, and saying so is part of the answer. At x=3x = 3: ln⁡122=ln⁡6≈1.7918\ln\frac{12}{2} = \ln 6 \approx 1.7918; directly, 2ln⁡4−ln⁡8+ln⁡3=ln⁡16⋅38=ln⁡62\ln 4 - \ln 8 + \ln 3 = \ln\frac{16 \cdot 3}{8} = \ln 6.

c) Quotient law, then product law, then power law, each legal because xx and yy are positive: log⁡2(8x3)−log⁡2y=log⁡28+log⁡2x3−log⁡2y1/2=3+3log⁡2x−12log⁡2y\log_2(8x^3) - \log_2 \sqrt y = \log_2 8 + \log_2 x^3 - \log_2 y^{1/2} = 3 + 3\log_2 x - \frac{1}{2}\log_2 y. Two points are lost here in practice: log⁡28\log_2 8 left unevaluated or written log⁡23\log_2 3, and y\sqrt y not rewritten as y1/2y^{1/2} before applying the power law, which produces log⁡2y\sqrt{\log_2 y}. Rewrite radicals as exponents FIRST.

d) (i) FALSE: a=b=1a = b = 1 gives ln⁡2≈0.69\ln 2 \approx 0.69 on the left and 0+0=00 + 0 = 0 on the right. (ii) FALSE: a=b=ea = b = e gives ln⁡e2=2\ln e^2 = 2 against 1⋅1=11 \cdot 1 = 1; the true law turns a product into a SUM. (iii) FALSE: a=8a = 8, b=2b = 2 gives ln⁡4≈1.386\ln 4 \approx 1.386 against ln⁡8ln⁡2=3\frac{\ln 8}{\ln 2} = 3; a quotient of logarithms is a change of base, ln⁡aln⁡b=log⁡ba\frac{\ln a}{\ln b} = \log_b a, never the logarithm of a quotient. (iv) TRUE: log⁡ab⋅log⁡ba=ln⁡bln⁡a⋅ln⁡aln⁡b=1\log_a b \cdot \log_b a = \frac{\ln b}{\ln a}\cdot\frac{\ln a}{\ln b} = 1. Each fake law comes from moving the logarithm like a common factor; a logarithm is a function, and ln⁡(a+b)\ln(a + b) has no simpler form at all.

e) Change of base with the ln⁡\ln key: log⁡320=ln⁡20ln⁡3≈2.9957321.098612≈2.7268\log_3 20 = \frac{\ln 20}{\ln 3} \approx \frac{2.995732}{1.098612} \approx 2.7268, and log⁡0.512=ln⁡12ln⁡0.5≈2.484907−0.693147≈−3.5850\log_{0.5} 12 = \frac{\ln 12}{\ln 0.5} \approx \frac{2.484907}{-0.693147} \approx -3.5850 (the log⁡\log key gives the same ratios). Before any key: 32=9<20<27=333^2 = 9 < 20 < 27 = 3^3 and log⁡3\log_3 is increasing, so 2<log⁡320<32 < \log_3 20 < 3. For the second, a base below 11 makes the logarithm DECREASING, so a number above 11 has a negative logarithm, as found: (0.5)−3.585=23.585≈12(0.5)^{-3.585} = 2^{3.585} \approx 12. A bracket by powers of the base takes five seconds and catches an inverted ratio, ln⁡3ln⁡20≈0.3667\frac{\ln 3}{\ln 20} \approx 0.3667, every time.

Exercise 5: e to the ln x against ln of e to the x: the cancellation laws and the domains they carry

exe^x and ln⁡x\ln x are inverse functions, which gives the two cancellation laws: ln⁡(ex)=x\ln(e^x) = x for EVERY real xx, and eln⁡x=xe^{\ln x} = x for x>0x > 0 ONLY, since ln⁡x\ln x does not exist otherwise. The same holds for any base: log⁡a(ax)=x\log_a(a^x) = x for all xx, alog⁡ax=xa^{\log_a x} = x for x>0x > 0.

The figure shows the dashed parabola y=x2y = x^2 and the graph of y=e2ln⁡xy = e^{2\ln x}, with an open circle at the origin.

-4-3-2-11234-224681012y = exp(2 ln x)y = x²
  • a) Simplify, stating for which xx each simplification is valid: eln⁡(x+2)e^{\ln(x + 2)}, ln⁡(e3x−1)\ln\left(e^{3x - 1}\right), e2ln⁡xe^{2\ln x}. Compute e−ln⁡4e^{-\ln 4} exactly.
  • b) Using the figure, is e2ln⁡xe^{2\ln x} defined at x=−3x = -3? Solve e2ln⁡x=9e^{2\ln x} = 9.
  • c) Find the domain of A(x)=ln⁡(1−ln⁡x)A(x) = \ln(1 - \ln x) and of B(x)=log⁡2(5−2x)B(x) = \log_2(5 - 2^x).
  • d) Solve eln⁡(x2−5)=4xe^{\ln(x^2 - 5)} = 4x.

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

a)
eln⁡(x+2)=x+2e^{\ln(x + 2)} = x + 2 holds for xx in ,
ln⁡(e3x−1)=3x−1\ln\left(e^{3x - 1}\right) = 3x - 1 holds for xx in ,
e2ln⁡x=x2e^{2\ln x} = x^2 holds for xx in ,
b)
c)
d)
Show the solution

Answers

  • a) x+2x + 2 for x>−2x > -2; 3x−13x - 1 for every xx; x2x^2 for x>0x > 0 only; e−ln⁡4=14e^{-\ln 4} = \frac{1}{4}
  • b) No; x=3x = 3 only (x=−3x = -3 rejected)
  • c) AA: 0<x<e0 < x < e; BB: x<log⁡25≈2.3219x < \log_2 5 \approx 2.3219
  • d) x=5x = 5 (x=−1x = -1 rejected)

a) eln⁡(x+2)=x+2e^{\ln(x + 2)} = x + 2, valid only where ln⁡(x+2)\ln(x + 2) exists, x>−2x > -2. ln⁡(e3x−1)=3x−1\ln\left(e^{3x - 1}\right) = 3x - 1 for every real xx, because e3x−1e^{3x-1} is always positive: this direction never costs a domain. e2ln⁡x=eln⁡x2=x2e^{2\ln x} = e^{\ln x^2} = x^2, by the power law, but only for x>0x > 0, since ln⁡x\ln x is the first thing computed. e−ln⁡4=eln⁡4−1=14e^{-\ln 4} = e^{\ln 4^{-1}} = \frac{1}{4}. The rule to remember: when ee is applied LAST, the domain of the ln⁡\ln inside survives the simplification, and the simplified formula must carry it.

b) No: at x=−3x = -3 the expression e2ln⁡(−3)e^{2\ln(-3)} requires ln⁡(−3)\ln(-3), which does not exist. That is what the figure shows: e2ln⁡xe^{2\ln x} is only the RIGHT half of the parabola, the open circle marking that x=0x = 0 is excluded too. Solving e2ln⁡x=9e^{2\ln x} = 9: on its domain x>0x > 0 the equation is x2=9x^2 = 9, whose roots are 33 and −3-3; the root −3-3 is outside the domain and is rejected. The only solution is x=3x = 3. A student who simplifies to x2=9x^2 = 9 and forgets where that simplification holds answers x=±3x = \pm 3 and loses half the marks.

c) A(x)=ln⁡(1−ln⁡x)A(x) = \ln(1 - \ln x) needs two conditions, one per logarithm: x>0x > 0 for the inner one, and 1−ln⁡x>01 - \ln x > 0 for the outer one, that is ln⁡x<1\ln x < 1, that is x<ex < e (ln⁡\ln is increasing, so the inequality keeps its direction when ee is applied). Domain: (0,e)(0, e), with e≈2.7183e \approx 2.7183. B(x)=log⁡2(5−2x)B(x) = \log_2(5 - 2^x) needs 2x<52^x < 5; apply the increasing function log⁡2\log_2: x<log⁡25=ln⁡5ln⁡2≈2.3219x < \log_2 5 = \frac{\ln 5}{\ln 2} \approx 2.3219. Domain: (−∞,log⁡25)(-\infty, \log_2 5). Forgetting the inner condition x>0x > 0 of AA is the usual slip: the inequality ln⁡x<1\ln x < 1 alone seems to allow x=−1x = -1, where nothing exists.

d) The left side exists only when x2−5>0x^2 - 5 > 0, that is x<−5x < -\sqrt 5 or x>5x > \sqrt 5. On that set, eln⁡(x2−5)=x2−5e^{\ln(x^2 - 5)} = x^2 - 5, and the equation becomes x2−5=4xx^2 - 5 = 4x, that is x2−4x−5=(x−5)(x+1)=0x^2 - 4x - 5 = (x - 5)(x + 1) = 0. Candidates: x=5x = 5 and x=−1x = -1. For x=5x = 5: x2−5=20>0x^2 - 5 = 20 > 0, and eln⁡20=20=4⋅5e^{\ln 20} = 20 = 4 \cdot 5, kept. For x=−1x = -1: x2−5=−4<0x^2 - 5 = -4 < 0, so ln⁡(−4)\ln(-4) does not exist: REJECTED, although it solves the simplified equation. The rejection must be written with its reason: the cancellation law eln⁡u=ue^{\ln u} = u holds for u>0u > 0 only.

Part B: problems and reasoning (/50)

Exercise 6: An inverse from start to finish: e to the x minus four e to the minus x

Let k(x)=ex−4e−xk(x) = e^x - 4e^{-x}, defined for every real xx. The figure shows its graph and three points with exact coordinates. Finding k−1k^{-1} is a typical final exam question: the unknown appears twice, in exe^x and in e−xe^{-x}, and the whole difficulty is to turn that into ONE unknown.

Use only the facts of the chapter: e−x=1exe^{-x} = \frac{1}{e^x}, ex>0e^x > 0, and ln⁡\ln undoes exe^x. A calculator is allowed for the decimal values.

-1-0.50.511.522.53-9-6-3369(0, -3)(ln 2, 0)(ln 4, 3)y = k(x)
  • a) Explain, without any derivative, why kk is increasing and therefore one-to-one. Compute k(0)k(0), k(ln⁡2)k(\ln 2) and k(ln⁡4)k(\ln 4) exactly.
  • b) Write y=k(x)y = k(x) and multiply both sides by exe^x. Show that u=exu = e^x satisfies u2−yu−4=0u^2 - yu - 4 = 0, solve for uu, and explain which root must be rejected, for every yy. Do it first with y=3y = 3.
  • c) Deduce a formula for k−1(x)k^{-1}(x) and its domain. Check it with k−1(0)k^{-1}(0) and k−1(−3)k^{-1}(-3).
  • d) Solve k(x)=6k(x) = 6 and k(x)=−6k(x) = -6 exactly, then to four decimals.
  • e) Show that k(ln⁡2+t)=−k(ln⁡2−t)k(\ln 2 + t) = -k(\ln 2 - t) for every tt, deduce that k−1(−y)=2ln⁡2−k−1(y)k^{-1}(-y) = 2\ln 2 - k^{-1}(y), and use it to find k−1(6)+k−1(−6)k^{-1}(6) + k^{-1}(-6) without d).

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

a)
b)
c)
Domain of k−1k^{-1} ,
d)
e)
Show the solution

Answers

  • a) exe^x and −4e−x-4e^{-x} both increase; k(0)=−3k(0) = -3, k(ln⁡2)=0k(\ln 2) = 0, k(ln⁡4)=3k(\ln 4) = 3
  • b) u=y±y2+162u = \frac{y \pm \sqrt{y^2 + 16}}{2}, the minus root is negative and rejected; for y=3y = 3: u=4u = 4 kept, u=−1u = -1 rejected
  • c) k−1(x)=ln⁡x+x2+162k^{-1}(x) = \ln\frac{x + \sqrt{x^2 + 16}}{2}, domain R\mathbb{R}; k−1(0)=ln⁡2k^{-1}(0) = \ln 2, k−1(−3)=0k^{-1}(-3) = 0
  • d) x=ln⁡(3+13)≈1.8879x = \ln(3 + \sqrt{13}) \approx 1.8879; x=ln⁡(13−3)≈−0.5016x = \ln(\sqrt{13} - 3) \approx -0.5016
  • e) k(ln⁡2+t)=2et−2e−tk(\ln 2 + t) = 2e^t - 2e^{-t}; k−1(6)+k−1(−6)=2ln⁡2=ln⁡4≈1.3863k^{-1}(6) + k^{-1}(-6) = 2\ln 2 = \ln 4 \approx 1.3863

a) When xx increases, exe^x increases and e−x=1exe^{-x} = \frac{1}{e^x} decreases, so −4e−x-4e^{-x} INCREASES. A sum of two increasing functions is increasing: if a<ba < b, then ea<ebe^a < e^b and −4e−a<−4e−b-4e^{-a} < -4e^{-b}, and adding gives k(a)<k(b)k(a) < k(b). An increasing function never takes a value twice, so kk is one-to-one. Values: k(0)=1−4=−3k(0) = 1 - 4 = -3; k(ln⁡2)=eln⁡2−4e−ln⁡2=2−4⋅12=0k(\ln 2) = e^{\ln 2} - 4e^{-\ln 2} = 2 - 4\cdot\frac{1}{2} = 0; k(ln⁡4)=4−4⋅14=3k(\ln 4) = 4 - 4\cdot\frac{1}{4} = 3. The step students get wrong is e−ln⁡2e^{-\ln 2}: it is 12\frac{1}{2}, the reciprocal, not −2-2.

b) y=ex−4e−xy = e^x - 4e^{-x}. Multiplying by exe^x, which is never 00, clears the negative exponent: yex=e2x−4ye^x = e^{2x} - 4, since e−x⋅ex=1e^{-x}\cdot e^x = 1. With u=exu = e^x: u2−yu−4=0u^2 - yu - 4 = 0, a quadratic in uu, and the two appearances of the unknown have become one. Quadratic formula: u=y±y2+162u = \frac{y \pm \sqrt{y^2 + 16}}{2}. Since y2+16>y2=∣y∣≥y\sqrt{y^2 + 16} > \sqrt{y^2} = |y| \ge y, the root with the minus sign is always NEGATIVE, and u=ex>0u = e^x > 0: it is rejected, for every yy, and that sentence is the mark. With y=3y = 3: u2−3u−4=(u−4)(u+1)=0u^2 - 3u - 4 = (u - 4)(u + 1) = 0, so u=4u = 4 is kept and u=−1u = -1 rejected; then ex=4e^x = 4 and x=ln⁡4x = \ln 4, which matches a).

c) Keep ex=y+y2+162e^x = \frac{y + \sqrt{y^2 + 16}}{2}, which is positive for every real yy, and apply ln⁡\ln: x=ln⁡y+y2+162x = \ln\frac{y + \sqrt{y^2 + 16}}{2}. Renaming, k−1(x)=ln⁡x+x2+162k^{-1}(x) = \ln\frac{x + \sqrt{x^2 + 16}}{2}, defined for EVERY real xx: the range of kk is all of R\mathbb{R}, as the figure suggests. Checks against a): k−1(0)=ln⁡0+42=ln⁡2k^{-1}(0) = \ln\frac{0 + 4}{2} = \ln 2, and k−1(−3)=ln⁡−3+52=ln⁡1=0k^{-1}(-3) = \ln\frac{-3 + 5}{2} = \ln 1 = 0. Writing ±\pm in the final formula would give two outputs for one input: an inverse function has one sign, chosen by the condition ex>0e^x > 0.

d) k(x)=6k(x) = 6 gives x=k−1(6)=ln⁡6+522=ln⁡6+2132=ln⁡(3+13)≈1.8879x = k^{-1}(6) = \ln\frac{6 + \sqrt{52}}{2} = \ln\frac{6 + 2\sqrt{13}}{2} = \ln(3 + \sqrt{13}) \approx 1.8879. The simplification 52=213\sqrt{52} = 2\sqrt{13} then the division of EVERY term by 22 is the algebra that is expected; dividing only the 66 gives ln⁡(3+213)\ln(3 + 2\sqrt{13}), a wrong value. Likewise k(x)=−6k(x) = -6 gives x=ln⁡−6+2132=ln⁡(13−3)≈ln⁡0.6056≈−0.5016x = \ln\frac{-6 + 2\sqrt{13}}{2} = \ln(\sqrt{13} - 3) \approx \ln 0.6056 \approx -0.5016, a negative number, consistent with k(0)=−3>−6k(0) = -3 > -6 and kk increasing. Check with the calculator: e1.8879−4e−1.8879≈6.606−0.606=6.000e^{1.8879} - 4e^{-1.8879} \approx 6.606 - 0.606 = 6.000.

e) k(ln⁡2+t)=2et−4⋅12e−t=2et−2e−tk(\ln 2 + t) = 2e^t - 4\cdot\frac{1}{2}e^{-t} = 2e^t - 2e^{-t}, and k(ln⁡2−t)=2e−t−2etk(\ln 2 - t) = 2e^{-t} - 2e^{t}, its opposite: the graph is symmetric about the point (ln⁡2,0)(\ln 2, 0). Now let a=k−1(y)a = k^{-1}(y) and t=a−ln⁡2t = a - \ln 2: then k(ln⁡2−t)=−k(ln⁡2+t)=−k(a)=−yk(\ln 2 - t) = -k(\ln 2 + t) = -k(a) = -y, so k−1(−y)=ln⁡2−t=2ln⁡2−ak^{-1}(-y) = \ln 2 - t = 2\ln 2 - a. Hence k−1(6)+k−1(−6)=2ln⁡2=ln⁡4≈1.3863k^{-1}(6) + k^{-1}(-6) = 2\ln 2 = \ln 4 \approx 1.3863, with no square root in sight. It agrees with d): ln⁡(3+13)+ln⁡(13−3)=ln⁡((13)2−32)=ln⁡4\ln(3 + \sqrt{13}) + \ln(\sqrt{13} - 3) = \ln\left((\sqrt{13})^2 - 3^2\right) = \ln 4, by the product law and a difference of squares.

Exercise 7: Exponential and logarithmic equations: collect, factor, then test every candidate

The method: isolate the exponential or the logarithm, apply the inverse function to both sides, solve, then TEST every candidate in the ORIGINAL equation. When the unknown appears in two exponents, take logarithms and collect the xx terms before dividing; when it appears as exe^x and e−xe^{-x}, multiply by exe^x first.

Give each answer exactly, then to four decimals with your calculator. A rejected candidate is named, with the reason.

  • a) Solve 32x+1=7x3^{2x + 1} = 7^x.
  • b) Solve ex+6e−x=5e^x + 6e^{-x} = 5.
  • c) Solve ln⁡(x+4)−ln⁡(x−2)=ln⁡x\ln(x + 4) - \ln(x - 2) = \ln x.
  • d) Solve log⁡2x+log⁡4x=6\log_2 x + \log_4 x = 6.
  • e) Solve (ln⁡x)2=ln⁡(x3)+4(\ln x)^2 = \ln\left(x^3\right) + 4.

Type your answers, the page tells you right or wrong 0/7

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

Answers

  • a) x=−ln⁡32ln⁡3−ln⁡7=−ln⁡3ln⁡(9/7)≈−4.3715x = \frac{-\ln 3}{2\ln 3 - \ln 7} = -\frac{\ln 3}{\ln(9/7)} \approx -4.3715
  • b) x=ln⁡2≈0.6931x = \ln 2 \approx 0.6931 or x=ln⁡3≈1.0986x = \ln 3 \approx 1.0986
  • c) x=4x = 4 (x=−1x = -1 rejected)
  • d) x=16x = 16
  • e) x=e4≈54.5982x = e^4 \approx 54.5982 or x=e−1≈0.3679x = e^{-1} \approx 0.3679

a) Different bases that are not powers of a common base: take ln⁡\ln of both sides, legal since both are positive. The power law gives (2x+1)ln⁡3=xln⁡7(2x + 1)\ln 3 = x\ln 7. Now the gesture: DISTRIBUTE, 2xln⁡3+ln⁡3=xln⁡72x\ln 3 + \ln 3 = x\ln 7, COLLECT the xx terms, 2xln⁡3−xln⁡7=−ln⁡32x\ln 3 - x\ln 7 = -\ln 3, FACTOR, x(2ln⁡3−ln⁡7)=−ln⁡3x(2\ln 3 - \ln 7) = -\ln 3, and divide: x=−ln⁡32ln⁡3−ln⁡7=−ln⁡3ln⁡(9/7)≈−1.0986120.251314≈−4.3715x = \frac{-\ln 3}{2\ln 3 - \ln 7} = -\frac{\ln 3}{\ln(9/7)} \approx -\frac{1.098612}{0.251314} \approx -4.3715. A negative answer is normal: 32x+13^{2x+1} grows faster than 7x7^x (base 99 against 77), so they meet on the left. Check: 32(−4.3715)+1≈2.03×10−43^{2(-4.3715) + 1} \approx 2.03 \times 10^{-4} and 7−4.3715≈2.03×10−47^{-4.3715} \approx 2.03 \times 10^{-4}. The error that costs the question is to divide by ln⁡3\ln 3 too early and write 2x+1=x⋅ln⁡7ln⁡32x + 1 = x \cdot \frac{\ln 7}{\ln 3}, then lose the xx on the right.

b) The unknown is in exe^x and in e−xe^{-x}: multiply both sides by ex>0e^x > 0, e2x+6=5exe^{2x} + 6 = 5e^x, and set u=exu = e^x: u2−5u+6=(u−2)(u−3)=0u^2 - 5u + 6 = (u - 2)(u - 3) = 0. Both roots are positive, so both are kept: ex=2e^x = 2 or ex=3e^x = 3, that is x=ln⁡2≈0.6931x = \ln 2 \approx 0.6931 or x=ln⁡3≈1.0986x = \ln 3 \approx 1.0986. Check: 2+62=52 + \frac{6}{2} = 5 and 3+63=53 + \frac{6}{3} = 5. Taking ln⁡\ln term by term, x+ln⁡6−x=ln⁡5x + \ln 6 - x = \ln 5, is the fake law ln⁡(a+b)=ln⁡a+ln⁡b\ln(a + b) = \ln a + \ln b and leads to the absurd ln⁡6=ln⁡5\ln 6 = \ln 5.

c) Domain of the ORIGINAL equation first: x+4>0x + 4 > 0, x−2>0x - 2 > 0 and x>0x > 0, so x>2x > 2. Quotient law: ln⁡x+4x−2=ln⁡x\ln\frac{x + 4}{x - 2} = \ln x, and ln⁡\ln is one-to-one, so x+4x−2=x\frac{x + 4}{x - 2} = x. Clear the denominator: x+4=x2−2xx + 4 = x^2 - 2x, so x2−3x−4=(x−4)(x+1)=0x^2 - 3x - 4 = (x - 4)(x + 1) = 0. Candidates 44 and −1-1. x=4>2x = 4 > 2 is kept: ln⁡8−ln⁡2=ln⁡4\ln 8 - \ln 2 = \ln 4. x=−1x = -1 is REJECTED because ln⁡(x−2)=ln⁡(−3)\ln(x - 2) = \ln(-3) does not exist, even though −1+4−1−2=−1\frac{-1 + 4}{-1 - 2} = -1 solves the rational equation. The quotient law and the clearing of the denominator both enlarged the domain; the test against x>2x > 2 is what restores it.

d) Two bases: bring them to one with the change of base, log⁡4x=log⁡2xlog⁡24=12log⁡2x\log_4 x = \frac{\log_2 x}{\log_2 4} = \frac{1}{2}\log_2 x. Then log⁡2x+12log⁡2x=32log⁡2x=6\log_2 x + \frac{1}{2}\log_2 x = \frac{3}{2}\log_2 x = 6, so log⁡2x=4\log_2 x = 4 and x=24=16x = 2^4 = 16, which is positive, so kept. Check: log⁡216+log⁡416=4+2=6\log_2 16 + \log_4 16 = 4 + 2 = 6. Adding the logarithms as if they had the same base, log⁡2x+log⁡4x=log⁡8x\log_2 x + \log_4 x = \log_8 x or log⁡6x\log_6 x, is invented algebra: the laws only combine logarithms of the SAME base.

e) Domain x>0x > 0. Rewrite ln⁡(x3)=3ln⁡x\ln(x^3) = 3\ln x (legal for x>0x > 0), but keep (ln⁡x)2(\ln x)^2 as it is: it is a square of a logarithm, not 2ln⁡x2\ln x. With t=ln⁡xt = \ln x: t2=3t+4t^2 = 3t + 4, t2−3t−4=(t−4)(t+1)=0t^2 - 3t - 4 = (t - 4)(t + 1) = 0, so ln⁡x=4\ln x = 4 or ln⁡x=−1\ln x = -1, that is x=e4≈54.5982x = e^4 \approx 54.5982 or x=e−1≈0.3679x = e^{-1} \approx 0.3679. Both are positive, so BOTH are kept: a negative value of t=ln⁡xt = \ln x is perfectly allowed, only xx itself must be positive. Replacing (ln⁡x)2(\ln x)^2 by 2ln⁡x2\ln x turns the equation into 2t=3t+42t = 3t + 4, one wrong root x=e−4x = e^{-4}, and the question is lost.

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, refute it with exact values, and write the correct statement.

  • a) (ln⁡x)2=2ln⁡x(\ln x)^2 = 2\ln x for every x>0x > 0.
  • b) log⁡2(−8)=−3\log_2(-8) = -3, because 23=82^3 = 8.
  • c) If ff is one-to-one, then f−1f^{-1} has the same domain as ff.
  • d) In a logarithmic equation, a negative candidate is always rejected.
  • e) The graph of f−1f^{-1} is the reflection of the graph of ff in the yy-axis.

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

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

Answers

  • a) False at x=ex = e: 1≠21 \ne 2. The law is ln⁡(x2)=2ln⁡x\ln(x^2) = 2\ln x for x>0x > 0.
  • b) False: 2y>02^y > 0, so log⁡2(−8)\log_2(-8) is undefined. It is log⁡218=−3\log_2 \frac{1}{8} = -3.
  • c) False: exe^x has domain R\mathbb{R}, ln⁡x\ln x has domain (0,∞)(0, \infty). Domain of f−1f^{-1} = range of ff.
  • d) False: log⁡4(x2)=log⁡4(3x+10)\log_4(x^2) = \log_4(3x + 10) keeps x=−2x = -2. Reject only a candidate that makes an argument ≤0\le 0.
  • e) False: exe^x reflected in the yy-axis is e−xe^{-x}, not ln⁡x\ln x. The mirror is the line y=xy = x.

a) FALSE. At x=ex = e: (ln⁡e)2=1(\ln e)^2 = 1 while 2ln⁡e=22\ln e = 2. At x=e2x = e^2 the two sides agree, 4=44 = 4, and at x=1x = 1 too, which is why a quick test can mislead; but agreeing at two points is not an identity. The power law moves an exponent that is INSIDE the logarithm: ln⁡(x2)=2ln⁡x\ln(x^2) = 2\ln x for x>0x > 0. In (ln⁡x)2(\ln x)^2 the square is outside, applied to the value of the logarithm, and nothing simplifies. This is the confusion that wrecked exercise 7 e).

b) FALSE. log⁡2(−8)=y\log_2(-8) = y would mean 2y=−82^y = -8, but 2y>02^y > 0 for every real yy: the logarithm of a negative number does not exist. The student mixed up two questions. log⁡28=3\log_2 8 = 3 because 23=82^3 = 8, and log⁡218=−3\log_2 \frac{1}{8} = -3 because 2−3=182^{-3} = \frac{1}{8}. Correct statement: a logarithm can be negative (its argument is then between 00 and 11), but its ARGUMENT must be positive.

c) FALSE. f(x)=exf(x) = e^x is one-to-one with domain R\mathbb{R}, but f−1(x)=ln⁡xf^{-1}(x) = \ln x has domain (0,∞)(0, \infty). Correct statement: the domain of f−1f^{-1} is the RANGE of ff, and the range of f−1f^{-1} is the domain of ff. The two domains coincide only in special cases such as f(x)=1xf(x) = \frac{1}{x} or f(x)=x3f(x) = x^3.

d) FALSE. Solve log⁡4(x2)=log⁡4(3x+10)\log_4(x^2) = \log_4(3x + 10). The logarithm is one-to-one, so x2=3x+10x^2 = 3x + 10, (x−5)(x+2)=0(x - 5)(x + 2) = 0. For x=−2x = -2: x2=4>0x^2 = 4 > 0 and 3x+10=4>03x + 10 = 4 > 0, both arguments exist, and log⁡44=log⁡44\log_4 4 = \log_4 4: x=−2x = -2 IS a solution, as is x=5x = 5. Correct statement: a candidate is rejected when it makes some argument of a logarithm zero or negative in the ORIGINAL equation, whatever its own sign. Rejecting by reflex costs a full solution.

e) FALSE. Reflecting in the yy-axis replaces xx by −x-x: from exe^x it produces e−xe^{-x}, a decreasing exponential, whereas the inverse of exe^x is ln⁡x\ln x, which is increasing and defined only for x>0x > 0. At x=1x = 1: e−1≈0.37e^{-1} \approx 0.37 while ln⁡1=0\ln 1 = 0. Correct statement: the graph of f−1f^{-1} is the reflection of the graph of ff in the line y=xy = x, because (a,b)(a, b) on ff becomes (b,a)(b, a) on f−1f^{-1}.

Exercise 9: Half-life of a medical tracer: the decay function and its inverse

A nuclear medicine unit receives a dose of technetium-99m whose activity is 200200 MBq (megabecquerels) at t=0t = 0. Its half-life is 66 hours: every 66 hours, the activity is divided by 22. So A(t)=200(12)t/6A(t) = 200\left(\frac{1}{2}\right)^{t/6}, with tt in hours. Knowing WHEN the activity reaches a given value means computing the inverse function, tt as a function of AA.

The figure shows AA with its values after one, two and three half-lives. Give exact answers, then decimals to two places with your calculator.

612182430362550751001251501752002251005025y = A(t)t (hours)activity (MBq)
  • a) Compute A(24)A(24) exactly and A(9)A(9) to two decimals. A student claims that after 99 hours, one and a half half-lives, the activity is 2003\frac{200}{3} MBq. Correct this.
  • b) Explain why AA is one-to-one on [0,∞)[0, \infty), and find its inverse t=A−1(y)t = A^{-1}(y) with its domain.
  • c) The dose is still usable for imaging while its activity is at least 3030 MBq. Until what time? When is the activity 5050 MBq?
  • d) For a second tracer, the activity is measured at 8080 MBq, then at 5050 MBq nine hours later. Find its half-life hh, exactly and to two decimals.
  • e) A tracer B, of half-life 1212 hours, starts at 5050 MBq at the same moment as the technetium dose. When do the two activities become equal, and what is that common activity?

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

a)
b)
Domain of A−1A^{-1} ,
c)
d)
e)
Show the solution

Answers

  • a) A(24)=12.5A(24) = 12.5 MBq; A(9)=20022=502≈70.71A(9) = \frac{200}{2\sqrt 2} = 50\sqrt 2 \approx 70.71 MBq, not 66.6766.67
  • b) AA is decreasing; A−1(y)=6log⁡2200y=6ln⁡(200/y)ln⁡2A^{-1}(y) = 6\log_2\frac{200}{y} = \frac{6\ln(200/y)}{\ln 2}, domain (0,200](0, 200]
  • c) until t=6log⁡2203≈16.42t = 6\log_2\frac{20}{3} \approx 16.42 h; 5050 MBq at t=12t = 12 h
  • d) h=9ln⁡2ln⁡(8/5)≈13.27h = \frac{9\ln 2}{\ln(8/5)} \approx 13.27 h
  • e) at t=24t = 24 h, both at 12.512.5 MBq

a) 24=4×624 = 4 \times 6 hours is four half-lives: A(24)=200(12)4=20016=12.5A(24) = 200\left(\frac{1}{2}\right)^4 = \frac{200}{16} = 12.5 MBq. For 99 hours, 96=32\frac{9}{6} = \frac{3}{2}: A(9)=200⋅2−3/2=20022=502≈70.71A(9) = 200 \cdot 2^{-3/2} = \frac{200}{2\sqrt 2} = 50\sqrt 2 \approx 70.71 MBq. The laws of exponents do the work: 2−3/2=123/2=1222^{-3/2} = \frac{1}{2^{3/2}} = \frac{1}{2\sqrt 2}. The student divided 200200 by 2×1.5=32 \times 1.5 = 3, as if the activity were divided by a quantity PROPORTIONAL to the time. An exponential divides by the same factor in equal times: after one half-life by 22, after half a half-life by 2\sqrt 2, so after 1.51.5 half-lives by 22≈2.832\sqrt 2 \approx 2.83, not by 33.

b) A(t)=200⋅btA(t) = 200 \cdot b^t with b=(12)1/6b = \left(\frac{1}{2}\right)^{1/6}, a base between 00 and 11, so AA is decreasing, hence one-to-one. On [0,∞)[0, \infty) its values fill (0,200](0, 200]: that is the domain of A−1A^{-1}. Solve y=200(12)t/6y = 200\left(\frac{1}{2}\right)^{t/6}: isolate the power, (12)t/6=y200\left(\frac{1}{2}\right)^{t/6} = \frac{y}{200}, that is 2t/6=200y2^{t/6} = \frac{200}{y}, then apply log⁡2\log_2: t6=log⁡2200y\frac{t}{6} = \log_2\frac{200}{y}. So A−1(y)=6log⁡2200y=6ln⁡(200/y)ln⁡2A^{-1}(y) = 6\log_2\frac{200}{y} = \frac{6\ln(200/y)}{\ln 2}, the second form being the one the calculator computes. Check: A−1(100)=6log⁡22=6A^{-1}(100) = 6\log_2 2 = 6, one half-life. Taking ln⁡\ln before dividing by 200200 produces ln⁡200+t6ln⁡12\ln 200 + \frac{t}{6}\ln\frac{1}{2}, correct but longer; dividing first is cleaner.

c) AA is decreasing, so A(t)≥30A(t) \ge 30 exactly when t≤A−1(30)=6log⁡220030=6log⁡2203=6ln⁡(20/3)ln⁡2≈6×1.8971200.693147≈16.42t \le A^{-1}(30) = 6\log_2\frac{200}{30} = 6\log_2\frac{20}{3} = \frac{6\ln(20/3)}{\ln 2} \approx \frac{6 \times 1.897120}{0.693147} \approx 16.42 hours. Sanity check: 203≈6.67\frac{20}{3} \approx 6.67 lies between 44 and 88, so the answer lies between two and three half-lives, 1212 and 1818 hours, which the figure confirms. For 5050 MBq: A−1(50)=6log⁡24=12A^{-1}(50) = 6\log_2 4 = 12 hours exactly, two half-lives, the marked point of the figure.

d) With half-life hh, the activity is multiplied by (12)9/h\left(\frac{1}{2}\right)^{9/h} in nine hours: 80(12)9/h=5080\left(\frac{1}{2}\right)^{9/h} = 50, so 29/h=8050=852^{9/h} = \frac{80}{50} = \frac{8}{5}. Take ln⁡\ln: 9hln⁡2=ln⁡85\frac{9}{h}\ln 2 = \ln\frac{8}{5}, so h=9ln⁡2ln⁡(8/5)≈9×0.6931470.470004≈13.27h = \frac{9\ln 2}{\ln(8/5)} \approx \frac{9 \times 0.693147}{0.470004} \approx 13.27 hours. Here the unknown is in the DENOMINATOR of the exponent: take the logarithm, then solve the linear equation in 1h\frac{1}{h}, and only then invert. Sanity check: 5050 is more than half of 8080, so less than one half-life has passed, and indeed 9<13.279 < 13.27.

e) Tracer B has activity 50(12)t/1250\left(\frac{1}{2}\right)^{t/12}. Equal activities: 200⋅2−t/6=50⋅2−t/12200 \cdot 2^{-t/6} = 50 \cdot 2^{-t/12}. Divide by 50⋅2−t/650 \cdot 2^{-t/6}, which gathers the unknown in ONE power: 4=2t/6−t/12=2t/124 = 2^{t/6 - t/12} = 2^{t/12}. So t12=2\frac{t}{12} = 2 and t=24t = 24 hours. The common activity is 200⋅2−4=50⋅2−2=12.5200 \cdot 2^{-4} = 50 \cdot 2^{-2} = 12.5 MBq; the solution figure shows the crossing. The faster decay of technetium eats its fourfold head start in 2424 hours. Taking ln⁡\ln of each side directly works too, ln⁡200−t6ln⁡2=ln⁡50−t12ln⁡2\ln 200 - \frac{t}{6}\ln 2 = \ln 50 - \frac{t}{12}\ln 2, and it is again collect and factor: tln⁡2(16−112)=ln⁡4t\ln 2\left(\frac{1}{6} - \frac{1}{12}\right) = \ln 4.

61218243036255075100125150175200225technetium, half-life 6 htracer B, half-life 12 h(24, 12.5)t (hours)activity (MBq)

Exercise 10: A final exam question: decibels and the Richter scale, logarithms as functions

The sound level of a sound of intensity II (in W/m²) is L(I)=10log⁡10II0L(I) = 10\log_{10}\frac{I}{I_0} decibels, with the reference intensity I0=10−12I_0 = 10^{-12} W/m². The figure shows the two scales side by side: each step of 1010 dB is a MULTIPLICATION of the intensity by 1010.

The magnitude of an earthquake is M=log⁡10AA0M = \log_{10}\frac{A}{A_0}, where AA is the amplitude recorded by a seismograph and A0A_0 a reference amplitude, and the energy it releases satisfies log⁡10E=4.8+1.5M\log_{10} E = 4.8 + 1.5M, with EE in joules. A calculator is allowed; round as asked.

L (dB)I (W/m²)010⁻¹²2010⁻¹⁰4010⁻⁸6010⁻⁶8010⁻⁴10010⁻²1201
  • a) Find the sound level of a sound of intensity 10−510^{-5} W/m², then of 3.2×10−43.2 \times 10^{-4} W/m², to two decimals.
  • b) Find the inverse function I(L)I(L). What intensity corresponds to the threshold of pain, 120120 dB?
  • c) One machine produces 8080 dB. Two identical machines run together, and intensities add. Find the new level to two decimals, and explain why the answer is not 160160 dB. By what factor is the intensity multiplied when the level rises by 1010 dB?
  • d) How many identical 8080 dB machines are needed for the total level to reach at least 9595 dB?
  • e) Compare an earthquake of magnitude 7.17.1 with one of magnitude 5.15.1: ratio of amplitudes. Then find the ratio of the energies of two earthquakes whose magnitudes differ by 11, to two decimals, and the magnitude of an earthquake that releases 10001000 times the energy of one of magnitude 5.15.1.

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

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

Answers

  • a) 7070 dB; 80+10log⁡1032≈85.0580 + 10\log_{10} 32 \approx 85.05 dB
  • b) I(L)=10−12⋅10L/10=10L/10−12I(L) = 10^{-12}\cdot 10^{L/10} = 10^{L/10 - 12}; I(120)=1I(120) = 1 W/m²
  • c) 80+10log⁡102≈83.0180 + 10\log_{10} 2 \approx 83.01 dB; levels do not add, intensities do; factor 1010
  • d) n≥101.5≈31.6n \ge 10^{1.5} \approx 31.6, so 3232 machines
  • e) 100100; 101.5≈31.6210^{1.5} \approx 31.62; magnitude 7.17.1

a) L(10−5)=10log⁡1010−510−12=10log⁡10107=70L(10^{-5}) = 10\log_{10}\frac{10^{-5}}{10^{-12}} = 10\log_{10} 10^{7} = 70 dB: the law of exponents first, 10−510−12=10−5+12\frac{10^{-5}}{10^{-12}} = 10^{-5 + 12}, then log⁡10107=7\log_{10} 10^7 = 7. For 3.2×10−43.2 \times 10^{-4}: II0=3.2×108\frac{I}{I_0} = 3.2 \times 10^{8}, and the product law splits it, log⁡10(3.2×108)=log⁡103.2+8≈0.50515+8\log_{10}(3.2 \times 10^8) = \log_{10} 3.2 + 8 \approx 0.50515 + 8, so L≈85.05L \approx 85.05 dB. Order check: the intensity lies between 10−410^{-4} and 10−310^{-3}, so the level lies between 8080 and 9090 dB. The sign error in 10−5−(−12)10^{-5 - (-12)} gives −170-170 dB, a level that makes no sense for a sound you can hear.

b) Solve L=10log⁡10II0L = 10\log_{10}\frac{I}{I_0} for II: divide by 1010, log⁡10II0=L10\log_{10}\frac{I}{I_0} = \frac{L}{10}, then apply the inverse function 10(⋅)10^{(\cdot)}: II0=10L/10\frac{I}{I_0} = 10^{L/10}, so I(L)=10−12⋅10L/10=10L/10−12I(L) = 10^{-12}\cdot 10^{L/10} = 10^{L/10 - 12} W/m². At the threshold of pain, I(120)=1012−12=1I(120) = 10^{12 - 12} = 1 W/m². The inverse is what the figure shows when read from top to bottom: 120120 on the decibel scale sits above 1=1001 = 10^0 on the intensity scale. Forgetting to divide by 1010 before exponentiating gives 10L−1210^{L - 12} and 1010810^{108} W/m², an absurd intensity.

c) One machine: I1=I(80)=10−4I_1 = I(80) = 10^{-4} W/m². Two machines: 2×10−42 \times 10^{-4}, so L=10log⁡102×10−410−12=10(log⁡102+8)=80+10log⁡102≈83.01L = 10\log_{10}\frac{2 \times 10^{-4}}{10^{-12}} = 10(\log_{10} 2 + 8) = 80 + 10\log_{10} 2 \approx 83.01 dB. Adding the levels, 80+80=16080 + 80 = 160 dB, is the fake law log⁡(a+b)=log⁡a+log⁡b\log(a + b) = \log a + \log b: a logarithm turns a PRODUCT into a sum, and doubling the intensity adds only 10log⁡102≈3.0110\log_{10} 2 \approx 3.01 dB. Rising by 1010 dB multiplies II0\frac{I}{I_0} by 1010/10=1010^{10/10} = 10, which is the reading of the figure: equal steps on the decibel scale are equal RATIOS of intensity.

d) nn machines give n×10−4n \times 10^{-4} W/m², so L=80+10log⁡10nL = 80 + 10\log_{10} n. We need 80+10log⁡10n≥9580 + 10\log_{10} n \ge 95, that is log⁡10n≥1.5\log_{10} n \ge 1.5, that is n≥101.5≈31.62n \ge 10^{1.5} \approx 31.62, since 10x10^x is increasing. The number of machines is a whole number, so 3232 machines; 3131 give only 80+10log⁡1031≈94.9180 + 10\log_{10} 31 \approx 94.91 dB. The rounding goes UP here, because the condition is at least 9595 dB: rounding 31.6231.62 to 3232 is not a matter of decimals, it is the smallest integer that satisfies the inequality.

e) M1−M2=log⁡10A1A0−log⁡10A2A0=log⁡10A1A2M_1 - M_2 = \log_{10}\frac{A_1}{A_0} - \log_{10}\frac{A_2}{A_0} = \log_{10}\frac{A_1}{A_2}, by the quotient law: the reference amplitude cancels. So A1A2=107.1−5.1=102=100\frac{A_1}{A_2} = 10^{7.1 - 5.1} = 10^2 = 100. For the energies, log⁡10E1−log⁡10E2=1.5(M1−M2)\log_{10} E_1 - \log_{10} E_2 = 1.5(M_1 - M_2), so a difference of one magnitude gives E1E2=101.5≈31.62\frac{E_1}{E_2} = 10^{1.5} \approx 31.62. Finally, 10001000 times the energy means 1.5 ΔM=log⁡101000=31.5\,\Delta M = \log_{10} 1000 = 3, so ΔM=2\Delta M = 2 and the magnitude is 5.1+2=7.15.1 + 2 = 7.1. The same pattern closes the whole chapter: to compare two values on a logarithmic scale, subtract the logarithms, then apply the inverse function to the difference.

© Ahmed Squalli Houssaini. Exercise set published at www.letuteurscientifique.ca/en/exercices/math203-inverse-functions-logarithms. 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