MATH 203 Calculus I • Concordia University, Montreal

Corrected exercises: Applied optimization (MATH 203)

This is the corrected exercise set for applied optimization in MATH 203, Differential and Integral Calculus I, at Concordia University, section 4.6 of Thomas' Calculus. It is the last chapter of the course and the one that uses all the others: a figure becomes a function, the product and chain rules give its derivative, and the extreme value results of sections 4.1 and 4.3 decide the answer. A scientific calculator is allowed, so answers that are cube roots or exponentials are given to the hundredth or to the cent, with the exact form next to them whenever it is natural.

The thread running through the set: in an optimization problem the derivative is one line, and the marks are lost in the ALGEBRA around it. Substitute the constraint and simplify before differentiating; then write f′f' as ONE FACTORED FRACTION. Its numerator gives the critical numbers, and the sign of its factors on the WHOLE domain gives the justification that the extremum is absolute, closed interval or not. Every exercise names the algebraic gesture where the points go: a negative exponent brought over a common denominator, a fraction split term by term, a common factor 2x2x factored instead of divided out, a radical isolated before squaring, an exponential factored because it is never zero, and a binomial square expanded in full, (h−9)2=h2−18h+81(h - 9)^2 = h^2 - 18h + 81.

The traps named in the solutions: the lost minus sign of ddxx−1\frac{d}{dx}x^{-1}, a fraction divided in one term only, a critical number deleted by dividing by xx, a root created by squaring, a+b\sqrt{a + b} written as a+b\sqrt a + \sqrt b, a critical number outside the domain when a cap or a short distance moves the optimum to an endpoint, a local maximum taken for the answer, the minimum of d2d^2 given as the distance, and a decimal rounded too early.

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

  • • Method (Thomas 4.6): read, draw, name the variables; write the constraint; write the objective in ONE variable; state its domain; find the critical numbers; justify the absolute extremum; answer the question asked, with units.
  • • Before differentiating: ax=ax−1\frac{a}{x} = ax^{-1}, whose derivative is −ax2-\frac{a}{x^2}; a fraction splits term by term, A−BC=AC−BC\frac{A - B}{C} = \frac{A}{C} - \frac{B}{C}, never AC−B\frac{A}{C} - B.
  • • After differentiating: write f′f' as one fraction, a−bx2=ax2−bx2a - \frac{b}{x^2} = \frac{ax^2 - b}{x^2}, and factor the numerator. Solve by factoring, never by dividing by an expression that can be 00; eue^{u} is never 00.
  • • Closed interval [a,b][a, b], ff continuous: compare ff at the critical numbers and at BOTH endpoints. Open or unbounded domain: the sign of f′f' on the whole domain (increasing then decreasing gives the absolute maximum).
  • • Radical equation: isolate the radical, square, then check every root in the unsquared equation. Distance: optimize D=d2D = d^2, then take the square root of the value.
  • • Classic shapes: open cup h=rh = r; open square-based box h=x2h = \frac{x}{2}; closed box a cube; Norman window h=rh = r; largest rectangle in a triangle: half its area; cone in a sphere of radius RR: h=4R3h = \frac{4R}{3}.

Part A: the basics (/50)

Exercise 1: A garden with two fence prices: a budget, then a fixed area

Thomas solves every applied optimization problem in the same order: read the problem, draw a picture, introduce variables, write the quantity to optimize as a function of ONE variable using the constraint, find its domain, then test the critical numbers AND justify that the extreme value found is absolute. The derivative is the easy line. The algebra before it and the justification after it are where the marks go.

A rectangular vegetable garden is fenced on its four sides. The front side, along the street, gets a decorative fence at 1818 dollars per metre; the back and the two sides get wire mesh at 66 dollars per metre. Let xx be the length of the front (and of the back) and yy the depth of the garden, in metres, as in the figure.

decorative front: 18 dollars per mstreetwire mesh: 6 dollars per mgardenxyy
  • a) The budget is 14401440 dollars. Write the cost constraint, the area AA as a function of xx alone, and the domain of xx.
  • b) Find the dimensions of the garden of largest area, and that area, with the Closed Interval Method.
  • c) Reverse problem: the garden must have an area of 18001800 m2^2 and cost as little as possible. Write the cost C(x)C(x), rewrite it with a negative exponent, differentiate, and write C′(x)C'(x) as ONE factored fraction. Justify that the critical number gives the absolute minimum on the whole domain.
  • d) A classmate writes, in c), C′(x)=24+21600x2C'(x) = 24 + \frac{21600}{x^2}, concludes that CC is increasing and that the cheapest garden has xx as small as possible. Find the algebra error, compute the correct C′(10)C'(10), and explain why the conclusion was absurd anyway.
  • e) Compare the answers of b) and c), and show that at the optimum of b) the budget is split exactly in half between the front and back on one side and the two sides on the other.

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

a)
Domain of xx ,
b)
c)
d)
e)
Show the solution

Answers

  • a) 24x+12y=144024x + 12y = 1440, so y=120−2xy = 120 - 2x; A(x)=120x−2x2A(x) = 120x - 2x^2 on [0,60][0, 60]
  • b) x=30x = 30 m, y=60y = 60 m, A=1800A = 1800 m2^2
  • c) C(x)=24x+21600x−1C(x) = 24x + 21600x^{-1}, C′(x)=24(x−30)(x+30)x2C'(x) = \frac{24(x - 30)(x + 30)}{x^2}: x=30x = 30 m, y=60y = 60 m, C=1440C = 1440 dollars
  • d) ddx(21600x−1)=−21600x−2\frac{d}{dx}\left(21600x^{-1}\right) = -21600x^{-2}; C′(10)=−192C'(10) = -192; C→∞C \to \infty as x→0+x \to 0^+
  • e) The same garden answers both; 24x=72024x = 720 dollars and 12y=72012y = 720 dollars.

a) The front and the back together cost 18x+6x=24x18x + 6x = 24x dollars, the two sides 2×6y=12y2 \times 6y = 12y dollars. The constraint is 24x+12y=144024x + 12y = 1440; dividing by 1212 first keeps the numbers small: 2x+y=1202x + y = 120, so y=120−2xy = 120 - 2x. The area is A=xy=x(120−2x)=120x−2x2A = xy = x(120 - 2x) = 120x - 2x^2. Domain: x≥0x \ge 0 and y=120−2x≥0y = 120 - 2x \ge 0, so x∈[0,60]x \in [0, 60]. Both endpoints are flat gardens of area 00; including them makes the interval closed, and AA, a polynomial, is continuous on it: the Closed Interval Method applies.

b) A′(x)=120−4xA'(x) = 120 - 4x, zero for x=30x = 30, inside [0,60][0, 60]. Values: A(0)=0A(0) = 0, A(60)=0A(60) = 0, A(30)=30×60=1800A(30) = 30 \times 60 = 1800. The largest garden is 3030 m along the street and y=120−60=60y = 120 - 60 = 60 m deep, with an area of 18001800 m2^2. The answer is the pair of dimensions AND the area: x=30x = 30 alone answers nothing.

c) Now the constraint is the area, xy=1800xy = 1800, so y=1800xy = \frac{1800}{x}, and the cost is C(x)=24x+12⋅1800x=24x+21600x−1C(x) = 24x + 12 \cdot \frac{1800}{x} = 24x + 21600x^{-1}, on the domain (0,∞)(0, \infty): any positive front works, and there is no endpoint. Power rule on the negative exponent: C′(x)=24−21600x−2=24−21600x2C'(x) = 24 - 21600x^{-2} = 24 - \frac{21600}{x^2}. Over the common denominator x2x^2: C′(x)=24x2−21600x2=24(x2−900)x2=24(x−30)(x+30)x2C'(x) = \frac{24x^2 - 21600}{x^2} = \frac{24(x^2 - 900)}{x^2} = \frac{24(x - 30)(x + 30)}{x^2}. This ONE factored fraction does all the work. On (0,∞)(0, \infty), x2>0x^2 > 0 and x+30>0x + 30 > 0, so the sign of C′C' is the sign of x−30x - 30: negative on (0,30)(0, 30), positive on (30,∞)(30, \infty). CC decreases then increases on the WHOLE domain, so C(30)C(30) is the absolute minimum (the root x=−30x = -30 is not in the domain). Then y=60y = 60 m and C(30)=720+720=1440C(30) = 720 + 720 = 1440 dollars.

d) The error is in the derivative of 21600x\frac{21600}{x}. Rewritten as 21600x−121600x^{-1}, the power rule gives 21600×(−1)x−2=−21600x221600 \times (-1)x^{-2} = -\frac{21600}{x^2}: the exponent −1-1 comes down as a factor and carries its MINUS sign. The classmate lost it. Correctly, C′(10)=24−21600100=24−216=−192<0C'(10) = 24 - \frac{21600}{100} = 24 - 216 = -192 < 0. The conclusion was absurd before any derivative: as x→0+x \to 0^+, 21600x→∞\frac{21600}{x} \to \infty, a garden 11 cm wide and 1818 km deep needs an enormous amount of mesh. A sign check at one convenient point, here x=10x = 10, catches the lost minus in five seconds.

e) Both problems give the same garden, 3030 m by 6060 m, with area 18001800 m2^2 and cost 14401440 dollars. It is not a coincidence: if a garden of area 18001800 could be fenced for less than 14401440 dollars, the leftover money would enlarge it beyond 18001800 m2^2, contradicting b). At the optimum, the front and back cost 24×30=72024 \times 30 = 720 dollars and the two sides 12×60=72012 \times 60 = 720 dollars: half of the budget each. In general, A=x⋅B−24x12A = x \cdot \frac{B - 24x}{12} has A′=B−48x12=0A' = \frac{B - 48x}{12} = 0 for 24x=B224x = \frac{B}{2}, whatever the budget BB.

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: The box with a square base and a fixed amount of cardboard

An open box (no lid) has a square base of side xx cm and a height of hh cm. It is made from exactly 27002700 cm2^2 of cardboard, with no waste: the figure shows its net, the base and four side flaps. We want the box of largest volume.

The constraint gives hh as a FRACTION of xx. How that fraction is substituted and simplified decides everything that follows.

basexxhhno lid: 2700 cm² of cardboard in all
  • a) Write the constraint, solve it for hh, and show that V(x)=675x−x34V(x) = 675x - \frac{x^3}{4}. Give the domain of xx and the height of the box when x=10x = 10.
  • b) Find the dimensions of the box of largest volume and that volume, justifying that the maximum is absolute.
  • c) A classmate simplifies V=x2⋅2700−x24xV = x^2 \cdot \frac{2700 - x^2}{4x} as V=675x−x2V = 675x - x^2. Which critical number does this wrong formula give, and why does that number alone prove that the algebra is wrong?
  • d) Same 27002700 cm2^2, but the box now has a lid (a closed box with square base). Find its largest volume, to the nearest cm3^3, and the shape of the optimal box.
  • e) For an open box with square base made of MM cm2^2 of material, prove that the optimal box always has h=x2h = \frac{x}{2}.

Type your answers, the page tells you right or wrong 0/10

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

Answers

  • a) x2+4xh=2700x^2 + 4xh = 2700, h=2700−x24xh = \frac{2700 - x^2}{4x}; V=x(2700−x2)4V = \frac{x(2700 - x^2)}{4} on [0,303][0, 30\sqrt 3]; h(10)=65h(10) = 65 cm
  • b) V′(x)=3(30−x)(30+x)4V'(x) = \frac{3(30 - x)(30 + x)}{4}: x=30x = 30 cm, h=15h = 15 cm, V=13 500V = 13\,500 cm3^3
  • c) x=337.5x = 337.5, outside the domain [0,303][0, 30\sqrt 3]: the 4x4x divides the WHOLE numerator
  • d) h=1350−x22xh = \frac{1350 - x^2}{2x}, x=h=152x = h = 15\sqrt 2 cm: a cube, V=67502≈9546V = 6750\sqrt 2 \approx 9546 cm3^3
  • e) V′=M−3x24V' = \frac{M - 3x^2}{4}, x2=M3x^2 = \frac{M}{3}, h=M−x24x=M6x=x2h = \frac{M - x^2}{4x} = \frac{M}{6x} = \frac{x}{2}

a) The base has area x2x^2 and each of the four flaps xhxh: the constraint is x2+4xh=2700x^2 + 4xh = 2700, so h=2700−x24xh = \frac{2700 - x^2}{4x}. Then V=x2h=x2⋅2700−x24xV = x^2h = x^2 \cdot \frac{2700 - x^2}{4x}. Cancel ONE factor xx between x2x^2 and 4x4x: V=x(2700−x2)4=2700x−x34=675x−x34V = \frac{x(2700 - x^2)}{4} = \frac{2700x - x^3}{4} = 675x - \frac{x^3}{4}. Domain: x>0x > 0 and h≥0h \ge 0, that is x2≤2700x^2 \le 2700, x≤2700=303≈51.96x \le \sqrt{2700} = 30\sqrt 3 \approx 51.96. The simplified formula is a polynomial, so it also makes sense at x=0x = 0 with value 00 (no box); including that endpoint gives the closed interval [0,303][0, 30\sqrt 3]. For x=10x = 10: h=2700−10040=65h = \frac{2700 - 100}{40} = 65 cm, a tall narrow box.

b) V′(x)=675−3x24=2700−3x24=3(900−x2)4=3(30−x)(30+x)4V'(x) = 675 - \frac{3x^2}{4} = \frac{2700 - 3x^2}{4} = \frac{3(900 - x^2)}{4} = \frac{3(30 - x)(30 + x)}{4}. In the domain the only zero is x=30x = 30 (−30-30 is rejected). Closed Interval Method: V(0)=0V(0) = 0; V(303)=0V(30\sqrt 3) = 0 since then h=0h = 0; V(30)=675×30−270004=20250−6750=13 500V(30) = 675 \times 30 - \frac{27000}{4} = 20250 - 6750 = 13\,500. The largest box has a base of 3030 cm by 3030 cm and a height h=2700−900120=15h = \frac{2700 - 900}{120} = 15 cm, for 13 50013\,500 cm3^3. Check the constraint: 900+4×30×15=900+1800=2700900 + 4 \times 30 \times 15 = 900 + 1800 = 2700.

c) With V=675x−x2V = 675x - x^2, V′=675−2x=0V' = 675 - 2x = 0 gives x=337.5x = 337.5 cm, far beyond 303≈5230\sqrt 3 \approx 52: that box would need a base of more than 1111 m2^2 from 0.270.27 m2^2 of cardboard. A candidate outside the domain is often the first sign of an algebra slip. The slip: the denominator 4x4x divides the WHOLE numerator 2700−x22700 - x^2, so x2⋅2700−x24x=2700x4−x34x^2 \cdot \frac{2700 - x^2}{4x} = \frac{2700x}{4} - \frac{x^3}{4}. The classmate divided the first term by 4x4x and forgot to divide the second. Splitting a fraction is legal only term by term, and every term keeps the denominator.

d) Now there are two squares: 2x2+4xh=27002x^2 + 4xh = 2700, so h=2700−2x24x=1350−x22xh = \frac{2700 - 2x^2}{4x} = \frac{1350 - x^2}{2x} and V=x2h=x(1350−x2)2=675x−x32V = x^2h = \frac{x(1350 - x^2)}{2} = 675x - \frac{x^3}{2} on [0,1350][0, \sqrt{1350}]. V′(x)=675−3x22=3(450−x2)2V'(x) = 675 - \frac{3x^2}{2} = \frac{3(450 - x^2)}{2}, zero for x=450=152≈21.21x = \sqrt{450} = 15\sqrt 2 \approx 21.21 cm. VV vanishes at both endpoints and is positive in between, so this is the absolute maximum. Then h=1350−450302=900302=302=152h = \frac{1350 - 450}{30\sqrt 2} = \frac{900}{30\sqrt 2} = \frac{30}{\sqrt 2} = 15\sqrt 2: the height equals the side, the optimal closed box is a CUBE. V=(152)3=3375×22=67502≈9546V = (15\sqrt 2)^3 = 3375 \times 2\sqrt 2 = 6750\sqrt 2 \approx 9546 cm3^3 (calculator, nearest cm3^3). The lid costs material, so the volume drops from 13 50013\,500 to about 95469546 cm3^3.

e) x2+4xh=Mx^2 + 4xh = M gives h=M−x24xh = \frac{M - x^2}{4x} and V=Mx−x34V = \frac{Mx - x^3}{4}. V′(x)=M−3x24V'(x) = \frac{M - 3x^2}{4}, positive then negative around x2=M3x^2 = \frac{M}{3}, with V=0V = 0 at both endpoints: the absolute maximum. There, M−x2=3x2−x2=2x2M - x^2 = 3x^2 - x^2 = 2x^2, so h=2x24x=x2h = \frac{2x^2}{4x} = \frac{x}{2}. Replacing MM by 3x23x^2 instead of computing decimals is the algebraic gesture that makes the shape appear. With M=2700M = 2700: x=30x = 30, h=15h = 15, as in b).

Exercise 3: The container of one litre: an open cup, then a can with waste

A cylindrical container of radius rr cm and height hh cm must hold 10001000 cm3^3 (one litre). We want to use as little material as possible. The domain of rr will be OPEN, so the Closed Interval Method is not available and the justification must come from the sign of the derivative on the whole domain.

Answers are asked to the hundredth when the exact value is a cube root: a scientific calculator is allowed.

rhopen cup2rdisk cut from a squarewaste
  • a) The container is an OPEN cup (a bottom, no top). Write the area of material AA in terms of rr and hh, the volume constraint, A(r)A(r) in one variable, and its domain.
  • b) Write A′(r)A'(r) as one factored fraction, find the radius of the cup using the least material (to the hundredth), justify the absolute minimum, and give the least area to the hundredth.
  • c) Show, exactly and without any decimal, that the optimal cup has h=rh = r.
  • d) A CLOSED can of 10001000 cm3^3 is made from sheet metal, and its top and bottom disks are punched from squares of side 2r2r: the corners are wasted but paid for. Minimize the metal used, and give the radius, the metal used and the ratio hr\frac{h}{r}.
  • e) In b), a classmate writes 2πr−2000r−2=02\pi r - 2000r^{-2} = 0, then 2πr⋅r−2=20002\pi r \cdot r^{-2} = 2000, so r=π1000r = \frac{\pi}{1000}. What should both sides have been multiplied by?

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

a)
Domain of rr ,
b)
c)
d)
e)
Show the solution

Answers

  • a) A=πr2+2πrhA = \pi r^2 + 2\pi rh, πr2h=1000\pi r^2h = 1000, A(r)=πr2+2000rA(r) = \pi r^2 + \frac{2000}{r} on (0,∞)(0, \infty)
  • b) A′(r)=2(πr3−1000)r2A'(r) = \frac{2(\pi r^3 - 1000)}{r^2}: r=10π3≈6.83r = \frac{10}{\sqrt[3]{\pi}} \approx 6.83 cm, A=3πr2≈439.38A = 3\pi r^2 \approx 439.38 cm2^2
  • c) h=1000πr2=πr3πr2=rh = \frac{1000}{\pi r^2} = \frac{\pi r^3}{\pi r^2} = r
  • d) A=8r2+2000rA = 8r^2 + \frac{2000}{r}, A′=16(r3−125)r2A' = \frac{16(r^3 - 125)}{r^2}: r=5r = 5 cm, 600600 cm2^2, hr=8π≈2.55\frac{h}{r} = \frac{8}{\pi} \approx 2.55
  • e) By r2r^2: 2πr3=20002\pi r^3 = 2000.

a) The cup has one disk of area πr2\pi r^2 and a side of area 2πrh2\pi rh: A=πr2+2πrhA = \pi r^2 + 2\pi rh. The constraint is πr2h=1000\pi r^2 h = 1000, so h=1000πr2h = \frac{1000}{\pi r^2} (solve for hh, never for rr). Substituting, 2πr⋅1000πr2=2000r2\pi r \cdot \frac{1000}{\pi r^2} = \frac{2000}{r}, so A(r)=πr2+2000rA(r) = \pi r^2 + \frac{2000}{r}. Every r>0r > 0 gives a cup, so the domain is (0,∞)(0, \infty), open at 00 and unbounded.

b) Rewrite 2000r=2000r−1\frac{2000}{r} = 2000r^{-1}: A′(r)=2πr−2000r−2=2πr3−2000r2=2(πr3−1000)r2A'(r) = 2\pi r - 2000r^{-2} = \frac{2\pi r^3 - 2000}{r^2} = \frac{2(\pi r^3 - 1000)}{r^2}. On (0,∞)(0, \infty), r2>0r^2 > 0, and πr3−1000\pi r^3 - 1000 is increasing, negative for small rr and positive for large rr, zero once, at r3=1000πr^3 = \frac{1000}{\pi}, r=10π3≈6.83r = \frac{10}{\sqrt[3]{\pi}} \approx 6.83 cm. So AA decreases then increases on the WHOLE domain: the absolute minimum is at r≈6.83r \approx 6.83 cm. There πr3=1000\pi r^3 = 1000, so 2000r=2πr3r=2πr2\frac{2000}{r} = \frac{2\pi r^3}{r} = 2\pi r^2 and A=πr2+2πr2=3πr2≈439.38A = \pi r^2 + 2\pi r^2 = 3\pi r^2 \approx 439.38 cm2^2. Keep the exact rr in the calculator for this last step: rounding rr to 6.836.83 first and then computing 3πr23\pi r^2 gives 439.66439.66, the wrong hundredth.

c) h=1000πr2h = \frac{1000}{\pi r^2}. At the optimum, 1000=πr31000 = \pi r^3: substitute it in the numerator, h=πr3πr2=rh = \frac{\pi r^3}{\pi r^2} = r. The open cup is exactly as tall as its radius, half as tall as it is wide. Using the equation A′(r)=0A'(r) = 0 to replace a number by an expression in rr is the algebraic gesture that turns a decimal answer into a shape.

d) Each disk now costs a square of area (2r)2=4r2(2r)^2 = 4r^2, so A=2×4r2+2πrh=8r2+2000rA = 2 \times 4r^2 + 2\pi rh = 8r^2 + \frac{2000}{r} on (0,∞)(0, \infty). A′(r)=16r−2000r2=16r3−2000r2=16(r3−125)r2A'(r) = 16r - \frac{2000}{r^2} = \frac{16r^3 - 2000}{r^2} = \frac{16(r^3 - 125)}{r^2}, negative on (0,5)(0, 5) and positive on (5,∞)(5, \infty): the absolute minimum is at r=5r = 5 cm. Metal used: A(5)=200+400=600A(5) = 200 + 400 = 600 cm2^2. Then h=100025π=40π≈12.73h = \frac{1000}{25\pi} = \frac{40}{\pi} \approx 12.73 cm and hr=8π≈2.55\frac{h}{r} = \frac{8}{\pi} \approx 2.55. Paying for the corners makes the ends expensive, so the optimal can is taller and narrower than a can with h=2rh = 2r.

e) To clear the negative exponent, multiply both sides of 2πr=2000r−22\pi r = 2000r^{-2} by r2r^2, since r−2⋅r2=1r^{-2} \cdot r^2 = 1: 2πr3=20002\pi r^3 = 2000, r3=1000πr^3 = \frac{1000}{\pi}. Multiplying by r−2r^{-2} instead makes the exponent MORE negative on the other side, and the result r=π1000≈0.003r = \frac{\pi}{1000} \approx 0.003 cm, a cup three hundredths of a millimetre wide, should have stopped the classmate. Better still: write A′A' as one fraction first, as in b), and the equation becomes numerator =0= 0.

Exercise 4: The closest point of a parabola: factor 2x, never divide by it

The distance from the origin to a point (x,y)(x, y) is d=x2+y2d = \sqrt{x^2 + y^2}. Since t↦tt \mapsto \sqrt t is increasing on [0,∞)[0, \infty), dd and its square D=d2D = d^2 are smallest (and largest) at the SAME points: we optimize DD, which has no square root, and take the square root of the VALUE only at the end.

The figure shows the parabola y=4−x2y = 4 - x^2, the origin OO and a point P(x,4−x2)P(x, 4 - x^2) of the curve.

-3-2-1123-112345y = 4 - x²P(x, 4 - x²)dOx
  • a) Show that D(x)=x4−7x2+16D(x) = x^4 - 7x^2 + 16, and compute D(1)D(1).
  • b) Compute D′(x)D'(x) and FACTOR it. List every critical number. A classmate divides 4x3−14x=04x^3 - 14x = 0 by xx: which critical number does the classmate lose?
  • c) Build the sign table of D′D' on R\mathbb{R}. Find the points of the parabola closest to the origin and the least distance, and say what x=0x = 0 is for DD.
  • d) Only the arch above the xx-axis is kept: −2≤x≤2-2 \le x \le 2. Find the point of the arch FARTHEST from the origin, and its distance. What does the classmate of b) answer?
  • e) The origin is replaced by the point (0,c)(0, c) of the yy-axis. For which values of cc is the vertex (0,4)(0, 4) the closest point of the parabola?

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

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

Answers

  • a) D=x2+(4−x2)2=x4−7x2+16D = x^2 + (4 - x^2)^2 = x^4 - 7x^2 + 16; D(1)=10D(1) = 10
  • b) D′(x)=2x(2x2−7)D'(x) = 2x(2x^2 - 7): x=0x = 0 and x=±72x = \pm\sqrt{\frac{7}{2}}; dividing by xx loses x=0x = 0
  • c) (±142,12)\left(\pm\frac{\sqrt{14}}{2}, \frac{1}{2}\right) at distance 152≈1.9365\frac{\sqrt{15}}{2} \approx 1.9365; x=0x = 0 is a local maximum of DD
  • d) The vertex (0,4)(0, 4), at distance 44; the classmate answers (±2,0)(\pm 2, 0), at distance 22
  • e) Dc′(x)=2x(2x2+2c−7)D_c'(x) = 2x(2x^2 + 2c - 7): the vertex is closest if and only if c≥72c \ge \frac{7}{2}

a) D=x2+y2D = x^2 + y^2 with y=4−x2y = 4 - x^2: D(x)=x2+(4−x2)2D(x) = x^2 + (4 - x^2)^2. Expand the square as a binomial, (4−x2)2=16−8x2+x4(4 - x^2)^2 = 16 - 8x^2 + x^4, never as 16−x416 - x^4 or 16+x416 + x^4: D(x)=x4−7x2+16D(x) = x^4 - 7x^2 + 16. The domain is R\mathbb{R}, since every xx gives a point of the parabola. D(1)=1−7+16=10D(1) = 1 - 7 + 16 = 10, which checks against the figure: P(1,3)P(1, 3) is at distance 10\sqrt{10} from OO.

b) D′(x)=4x3−14xD'(x) = 4x^3 - 14x. Factor the common factor 2x2x: D′(x)=2x(2x2−7)D'(x) = 2x(2x^2 - 7). A product is zero when one factor is zero: x=0x = 0, or x2=72x^2 = \frac{7}{2}, x=±72=±142≈±1.8708x = \pm\sqrt{\frac{7}{2}} = \pm\frac{\sqrt{14}}{2} \approx \pm 1.8708. Three critical numbers. Dividing both sides of 4x3−14x=04x^3 - 14x = 0 by xx is only legal when x≠0x \ne 0, and it silently deletes the solution x=0x = 0. The rule: move everything to one side and FACTOR; never divide an equation by an expression that can be zero.

c) Signs of the factors: 2x2x changes sign at 00, and 2x2−72x^2 - 7 is negative between −72-\sqrt{\frac{7}{2}} and 72\sqrt{\frac{7}{2}}, positive outside. So D′<0D' < 0 on (−∞,−72)\left(-\infty, -\sqrt{\frac{7}{2}}\right), D′>0D' > 0 on (−72,0)\left(-\sqrt{\frac{7}{2}}, 0\right), D′<0D' < 0 on (0,72)\left(0, \sqrt{\frac{7}{2}}\right) and D′>0D' > 0 on (72,∞)\left(\sqrt{\frac{7}{2}}, \infty\right). DD decreases, increases, decreases, increases, and D→∞D \to \infty as x→±∞x \to \pm\infty: the absolute minimum is the smaller of the two local minima, and they are equal by symmetry, D(±72)=494−492+16=154D\left(\pm\sqrt{\frac{7}{2}}\right) = \frac{49}{4} - \frac{49}{2} + 16 = \frac{15}{4}. The closest points are (±142,12)\left(\pm\frac{\sqrt{14}}{2}, \frac{1}{2}\right), since y=4−72=12y = 4 - \frac{7}{2} = \frac{1}{2}, and the least distance is 154=152≈1.9365\sqrt{\frac{15}{4}} = \frac{\sqrt{15}}{2} \approx 1.9365. At x=0x = 0, D′D' goes from positive to negative: by the First Derivative Test it is a local MAXIMUM of DD, the vertex at distance 44. A critical number is a candidate, not an answer.

d) On the closed interval [−2,2][-2, 2], the continuous DD reaches a maximum, and the Closed Interval Method compares ALL the critical numbers with the endpoints: D(−2)=16−28+16=4D(-2) = 16 - 28 + 16 = 4, D(±72)=154D\left(\pm\sqrt{\frac{7}{2}}\right) = \frac{15}{4}, D(0)=16D(0) = 16, D(2)=4D(2) = 4. The farthest point of the arch is the vertex (0,4)(0, 4), at distance 16=4\sqrt{16} = 4. The classmate of b), who lost x=0x = 0, compares only 154\frac{15}{4} and 44 and answers the endpoints (±2,0)(\pm 2, 0), at distance 22: half the true answer, lost in one illegal division.

e) Dc(x)=x2+(4−c−x2)2D_c(x) = x^2 + (4 - c - x^2)^2. Chain rule on the square, inner function 4−c−x24 - c - x^2: Dc′(x)=2x+2(4−c−x2)(−2x)=2x[1−2(4−c−x2)]=2x(2x2+2c−7)D_c'(x) = 2x + 2(4 - c - x^2)(-2x) = 2x\left[1 - 2(4 - c - x^2)\right] = 2x(2x^2 + 2c - 7). If c≥72c \ge \frac{7}{2}, then 2x2+2c−7≥2x2>02x^2 + 2c - 7 \ge 2x^2 > 0 for x≠0x \ne 0, so Dc′D_c' has the sign of xx: DcD_c decreases then increases, and the vertex is the closest point. If c<72c < \frac{7}{2}, the second factor vanishes at x=±7−2c2x = \pm\sqrt{\frac{7 - 2c}{2}} and, as in c), x=0x = 0 becomes a local maximum. So the vertex is closest exactly when c≥72c \ge \frac{7}{2}; c=0c = 0 gives back b).

Exercise 5: The rectangle in a right triangle: in the corner, then on the hypotenuse

A right triangle has legs of 1212 and 99 (so a hypotenuse of 1515), placed with its right angle at the origin, its legs along the axes and its hypotenuse on the line through (12,0)(12, 0) and (0,9)(0, 9). We cut from it the rectangle of largest area, in two different positions shown in the figure.

The constraint here is a similar-triangle ratio. Written with fractions, it is exact; cleared of its fractions early, it keeps the derivative a one line computation.

129x by y(x, y)129corner rectanglerectangle on the hypotenuse
  • a) Corner rectangle: two sides along the legs and the opposite corner (x,y)(x, y) on the hypotenuse. Show that y=9−3x4y = 9 - \frac{3x}{4}, write the area A(x)A(x) and its domain.
  • b) Find the corner rectangle of largest area and compare it with the area of the triangle.
  • c) Rectangle on the hypotenuse: one side lies on the hypotenuse and the two other corners are on the legs. Compute the altitude HH of the triangle relative to the hypotenuse, express the width ww of the rectangle in terms of its height tt by similar triangles, and find the largest rectangle of this kind.
  • d) Explain why b) and c) give the same area: prove that for ANY triangle of base bb and height HH, the largest rectangle standing on the base has half the area of the triangle.

Type your answers, the page tells you right or wrong 0/10

a)
Domain of xx ,
b)
c)
d)
Show the solution

Answers

  • a) 9−yx=912\frac{9 - y}{x} = \frac{9}{12}, so y=9−3x4y = 9 - \frac{3x}{4}; A(x)=9x−3x24A(x) = 9x - \frac{3x^2}{4} on [0,12][0, 12]
  • b) x=6x = 6, y=92y = \frac{9}{2}, A=27A = 27, half of the triangle's 5454
  • c) H=365=7.2H = \frac{36}{5} = 7.2; w=15−25t12w = 15 - \frac{25t}{12}; t=3.6t = 3.6, w=7.5w = 7.5, A=27A = 27
  • d) A(t)=bH(Ht−t2)A(t) = \frac{b}{H}(Ht - t^2), maximum at t=H2t = \frac{H}{2}: A=bH4A = \frac{bH}{4}, half of bH2\frac{bH}{2}

a) Above the rectangle sits a small right triangle with legs xx (horizontal) and 9−y9 - y (vertical), similar to the big one, whose legs are 1212 and 99. Corresponding sides: 9−yx=912=34\frac{9 - y}{x} = \frac{9}{12} = \frac{3}{4}, so 9−y=3x49 - y = \frac{3x}{4} and y=9−3x4y = 9 - \frac{3x}{4}, which is also the equation of the hypotenuse. Then A(x)=xy=9x−3x24A(x) = xy = 9x - \frac{3x^2}{4}, on [0,12][0, 12] (at both ends the rectangle is flat). The ratio must pair a vertical side with a vertical side: writing 9−yx=129\frac{9 - y}{x} = \frac{12}{9} gives y=9−4x3y = 9 - \frac{4x}{3}, which reaches 00 at x=6.75x = 6.75 instead of 1212, a check that fails at once.

b) A′(x)=9−3x2=3(6−x)2A'(x) = 9 - \frac{3x}{2} = \frac{3(6 - x)}{2}, zero at x=6x = 6. Closed Interval Method: A(0)=A(12)=0A(0) = A(12) = 0 and A(6)=54−27=27A(6) = 54 - 27 = 27. The largest corner rectangle is 66 by 92\frac{9}{2}, area 2727: exactly half of the triangle, whose area is 12×92=54\frac{12 \times 9}{2} = 54. Its corner is the midpoint of the hypotenuse.

c) Area of the triangle two ways: 12×15×H=54\frac{1}{2} \times 15 \times H = 54, so H=10815=365=7.2H = \frac{108}{15} = \frac{36}{5} = 7.2. The side of the rectangle opposite the hypotenuse is parallel to it, at distance 7.2−t7.2 - t from the right angle: it cuts off a small triangle similar to the big one in the ratio 7.2−t7.2\frac{7.2 - t}{7.2}. Hence w=15⋅7.2−t7.2=15−15t7.2w = 15 \cdot \frac{7.2 - t}{7.2} = 15 - \frac{15t}{7.2}. Clear the decimal: 157.2=15072=2512\frac{15}{7.2} = \frac{150}{72} = \frac{25}{12}, so w=15−25t12w = 15 - \frac{25t}{12} and A(t)=tw=15t−25t212A(t) = tw = 15t - \frac{25t^2}{12} on [0,7.2][0, 7.2]. A′(t)=15−25t6=0A'(t) = 15 - \frac{25t}{6} = 0 for t=9025=3.6t = \frac{90}{25} = 3.6, with A(0)=A(7.2)=0A(0) = A(7.2) = 0. So t=3.6t = 3.6, w=15−7.5=7.5w = 15 - 7.5 = 7.5 and A=3.6×7.5=27A = 3.6 \times 7.5 = 27. The figure's rectangle has its far corners at (0,4.5)(0, 4.5) and (6,0)(6, 0): 36+20.25=7.5\sqrt{36 + 20.25} = 7.5, as computed.

d) Triangle of base bb and height HH, rectangle of height tt standing on the base. Its top side cuts off a triangle similar to the whole in the ratio H−tH\frac{H - t}{H}, so w=b(H−t)Hw = \frac{b(H - t)}{H} and A(t)=bH(Ht−t2)A(t) = \frac{b}{H}(Ht - t^2) on [0,H][0, H]. A′(t)=bH(H−2t)=0A'(t) = \frac{b}{H}(H - 2t) = 0 for t=H2t = \frac{H}{2}, with A=0A = 0 at both ends: the maximum is bH⋅H24=bH4\frac{b}{H} \cdot \frac{H^2}{4} = \frac{bH}{4}, half of the triangle's bH2\frac{bH}{2}. In b), the base is the leg 1212 and H=9H = 9: t=4.5t = 4.5; in c), the base is the hypotenuse 1515 and H=7.2H = 7.2: t=3.6t = 3.6. Both give 542=27\frac{54}{2} = 27.

Part B: problems and reasoning (/50)

Exercise 6: The Norman window of fixed area: the least frame

A Norman window is a rectangle 2r2r wide and hh high, surmounted by a half-disk of radius rr whose diameter is the top side of the rectangle. This window must let in light through an area of exactly 22 m2^2, and we want the FRAME, the whole outline, to be as short as possible.

Here the area is the constraint and the perimeter the objective. The fraction that gives hh must be SPLIT before it is used, and the domain is open at one end and closed at the other.

2rhrarea 2 m²curved frame
  • a) Write the area constraint and show that h=1r−πr4h = \frac{1}{r} - \frac{\pi r}{4}. Deduce the length of the frame P(r)=4+π2r+2rP(r) = \frac{4 + \pi}{2}r + \frac{2}{r} and the domain of rr. Give hh when r=0.5r = 0.5 m, to four decimals.
  • b) Write P′(r)P'(r) as one fraction and find the critical number, to four decimals. Justify that it gives the absolute minimum of PP on its domain.
  • c) Show that at the optimum h=rh = r exactly, and that the least frame is 24+π2\sqrt{4 + \pi} m. Compare with the window reduced to a half-disk.
  • d) The curved part of the frame costs three times as much per metre as the straight parts. Find the radius of the cheapest frame, to four decimals, and prove that then hr=1+π\frac{h}{r} = 1 + \pi.
  • e) In one sentence: how does a dearer arc change the shape of the optimal window?

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

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

Answers

  • a) 2rh+πr22=22rh + \frac{\pi r^2}{2} = 2; P(r)=4+π2r+2rP(r) = \frac{4 + \pi}{2}r + \frac{2}{r} on (0,2π]\left(0, \frac{2}{\sqrt\pi}\right]; h(0.5)=2−π8≈1.6073h(0.5) = 2 - \frac{\pi}{8} \approx 1.6073 m
  • b) P′(r)=(4+π)r2−42r2P'(r) = \frac{(4 + \pi)r^2 - 4}{2r^2}: r=24+π≈0.7484r = \frac{2}{\sqrt{4 + \pi}} \approx 0.7484 m, minimum by the sign of P′P'
  • c) h=rh = r; Pmin⁡=24+π≈5.3448P_{\min} = 2\sqrt{4 + \pi} \approx 5.3448 m, less than 2(2+π)π≈5.8017\frac{2(2 + \pi)}{\sqrt\pi} \approx 5.8017 m for the half-disk
  • d) K(r)=4+5π2r+2rK(r) = \frac{4 + 5\pi}{2}r + \frac{2}{r}: r=24+5π≈0.4505r = \frac{2}{\sqrt{4 + 5\pi}} \approx 0.4505 m, hr=1+π\frac{h}{r} = 1 + \pi
  • e) The arc shrinks: the window becomes narrower and taller.

a) The light area is the rectangle plus the half-disk: 2rh+πr22=22rh + \frac{\pi r^2}{2} = 2, so h=2−πr222rh = \frac{2 - \frac{\pi r^2}{2}}{2r}. Split the fraction term by term, each term keeping the denominator 2r2r: h=22r−πr22⋅2r=1r−πr4h = \frac{2}{2r} - \frac{\pi r^2}{2 \cdot 2r} = \frac{1}{r} - \frac{\pi r}{4}. The frame is the bottom 2r2r, the two sides 2h2h and the half-circle πr\pi r: P=2r+πr+2h=2r+πr+2r−πr2=4+π2r+2rP = 2r + \pi r + 2h = 2r + \pi r + \frac{2}{r} - \frac{\pi r}{2} = \frac{4 + \pi}{2}r + \frac{2}{r}. Domain: r>0r > 0 and h≥0h \ge 0, that is πr22≤2\frac{\pi r^2}{2} \le 2, r≤2π≈1.1284r \le \frac{2}{\sqrt\pi} \approx 1.1284; at that end, h=0h = 0 and the window is a half-disk alone, a real window. So r∈(0,2π]r \in \left(0, \frac{2}{\sqrt\pi}\right], open at 00. For r=0.5r = 0.5: h=2−π8≈1.6073h = 2 - \frac{\pi}{8} \approx 1.6073 m.

b) P′(r)=4+π2−2r2=(4+π)r2−42r2P'(r) = \frac{4 + \pi}{2} - \frac{2}{r^2} = \frac{(4 + \pi)r^2 - 4}{2r^2}. The denominator is positive; the numerator is increasing in r>0r > 0 and vanishes for r2=44+πr^2 = \frac{4}{4 + \pi}, r=24+π≈0.7484r = \frac{2}{\sqrt{4 + \pi}} \approx 0.7484 m, which lies in the domain because 4+π>π4 + \pi > \pi. So P′<0P' < 0 before it and P′>0P' > 0 after it, on the WHOLE domain: PP decreases then increases, and this critical number gives the absolute minimum. No endpoint value is needed for the justification, and there is none at 00 anyway: P→∞P \to \infty as r→0+r \to 0^+.

c) At the optimum, (4+π)r2=4(4 + \pi)r^2 = 4, so 1r=(4+π)r4\frac{1}{r} = \frac{(4 + \pi)r}{4} and h=(4+π)r4−πr4=rh = \frac{(4 + \pi)r}{4} - \frac{\pi r}{4} = r. The frame: 2r=(4+π)r2\frac{2}{r} = \frac{(4 + \pi)r}{2}, so P=(4+π)r=(4+π)⋅24+π=24+π≈5.3448P = (4 + \pi)r = (4 + \pi) \cdot \frac{2}{\sqrt{4 + \pi}} = 2\sqrt{4 + \pi} \approx 5.3448 m. The half-disk alone needs P(2π)=(2+π)2π≈5.8017P\left(\frac{2}{\sqrt\pi}\right) = (2 + \pi)\frac{2}{\sqrt\pi} \approx 5.8017 m, more, as the sign of P′P' predicted. The same shape, h=rh = r, as the window of largest area for a given frame: fixing one quantity and optimizing the other gives the same optimal shape.

d) In units of the price of one metre of straight frame, the cost is K=2r+2h+3πr=2r+3πr+2r−πr2=4+5π2r+2rK = 2r + 2h + 3\pi r = 2r + 3\pi r + \frac{2}{r} - \frac{\pi r}{2} = \frac{4 + 5\pi}{2}r + \frac{2}{r}, on the same domain. K′(r)=(4+5π)r2−42r2K'(r) = \frac{(4 + 5\pi)r^2 - 4}{2r^2}, negative then positive around r=24+5π≈0.4505r = \frac{2}{\sqrt{4 + 5\pi}} \approx 0.4505 m, which is in the domain: the absolute minimum. There, 1r2=4+5π4\frac{1}{r^2} = \frac{4 + 5\pi}{4}, and hr=1r2−π4=4+5π−π4=1+π≈4.14\frac{h}{r} = \frac{1}{r^2} - \frac{\pi}{4} = \frac{4 + 5\pi - \pi}{4} = 1 + \pi \approx 4.14.

e) The dearer arc makes the optimal window narrower and much taller (hh goes from rr to about 4.14r4.14r): the design moves area from the expensive half-disk to the rectangle, whose frame is cheap. The method did not change at all: one factored fraction, its sign on the whole domain, then the quantities asked.

Exercise 7: A pipeline across a river: isolate the radical, square, check

A pumping station SS stands on one bank of a straight river 1.21.2 km wide. A factory FF stands on the other bank, 44 km downstream from the point CC directly across from SS. The pipeline runs under water in a straight line from SS to a point PP of the far bank, xx km downstream from CC, then along the bank from PP to FF, as in the figure.

Laying pipe under water costs 1.31.3 million dollars per km; along the bank, 0.50.5 million dollars per km. Costs are in millions of dollars. The equation C′(x)=0C'(x) = 0 contains a square root: the algebra of that equation is the whole difficulty.

SCPF1.2 kmriveralong the bank to F: 4 - x kmx
  • a) Write the total cost C(x)C(x) and its domain, and compute C(0)C(0).
  • b) Compute C′(x)C'(x), naming the inner function of the chain rule. Solve C′(x)=0C'(x) = 0 by isolating the radical, squaring, and checking every root obtained.
  • c) Find the cheapest route with the Closed Interval Method, and the saving compared with crossing straight to CC, in dollars.
  • d) Another factory stands only 0.40.4 km downstream from CC. Find the cheapest route to it and its cost, to four decimals.
  • e) For a river of width ww (the factory still far enough), show that the landing point is x=5w12x = \frac{5w}{12}. Where does the pipe land if the river is 2.42.4 km wide?

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

a)
Domain of xx ,
b)
c)
d)
e)
Show the solution

Answers

  • a) C(x)=1.31.44+x2+0.5(4−x)C(x) = 1.3\sqrt{1.44 + x^2} + 0.5(4 - x) on [0,4][0, 4]; C(0)=3.56C(0) = 3.56
  • b) C′(x)=1.3x1.44+x2−0.5C'(x) = \frac{1.3x}{\sqrt{1.44 + x^2}} - 0.5; squaring gives x=±0.5x = \pm 0.5, only x=0.5x = 0.5 is kept
  • c) C(0)=3.56C(0) = 3.56, C(0.5)=3.44C(0.5) = 3.44, C(4)≈5.429C(4) \approx 5.429: land 0.50.5 km downstream, 3.443.44 million dollars, saving 120 000120\,000 dollars
  • d) C′<0C' < 0 on [0,0.4][0, 0.4]: straight to the factory under water, 1.31.6≈1.64441.3\sqrt{1.6} \approx 1.6444 million dollars
  • e) 1.44x2=0.25w21.44x^2 = 0.25w^2, x=5w12x = \frac{5w}{12}; for w=2.4w = 2.4: x=1x = 1 km

a) Under water the pipe covers SP=1.22+x2=1.44+x2SP = \sqrt{1.2^2 + x^2} = \sqrt{1.44 + x^2} km (Pythagoras in the right triangle SCPSCP), along the bank PF=4−xPF = 4 - x km. So C(x)=1.31.44+x2+0.5(4−x)C(x) = 1.3\sqrt{1.44 + x^2} + 0.5(4 - x), for x∈[0,4]x \in [0, 4]: landing upstream of CC or beyond FF only adds pipe. C(0)=1.3×1.2+0.5×4=1.56+2=3.56C(0) = 1.3 \times 1.2 + 0.5 \times 4 = 1.56 + 2 = 3.56 million dollars.

b) Chain rule, inner function u=1.44+x2u = 1.44 + x^2, u′=2xu' = 2x: ddxu=u′2u=x1.44+x2\frac{d}{dx}\sqrt{u} = \frac{u'}{2\sqrt u} = \frac{x}{\sqrt{1.44 + x^2}}. So C′(x)=1.3x1.44+x2−0.5C'(x) = \frac{1.3x}{\sqrt{1.44 + x^2}} - 0.5, defined on the whole domain. Isolate the radical before squaring: 1.3x=0.51.44+x21.3x = 0.5\sqrt{1.44 + x^2}. Now square both sides: 1.69x2=0.25(1.44+x2)=0.36+0.25x21.69x^2 = 0.25(1.44 + x^2) = 0.36 + 0.25x^2, so 1.44x2=0.361.44x^2 = 0.36, x2=0.25x^2 = 0.25, x=±0.5x = \pm 0.5. Check each root in the UNSQUARED equation: for x=0.5x = 0.5, 1.3×0.5=0.651.3 \times 0.5 = 0.65 and 0.51.69=0.5×1.3=0.650.5\sqrt{1.69} = 0.5 \times 1.3 = 0.65, true; for x=−0.5x = -0.5, the left side is −0.65<0-0.65 < 0 and the right side positive: squaring CREATED this root, and it is outside the domain anyway. Two classic slips cost the question here: squaring the equation before isolating the radical (a cross term appears and nothing simplifies), and writing 1.44+x2=1.2+x\sqrt{1.44 + x^2} = 1.2 + x, which is false (at x=0.5x = 0.5: 1.31.3 against 1.71.7).

c) CC is continuous on the closed interval [0,4][0, 4]. C(0)=3.56C(0) = 3.56; C(0.5)=1.3×1.3+0.5×3.5=1.69+1.75=3.44C(0.5) = 1.3 \times 1.3 + 0.5 \times 3.5 = 1.69 + 1.75 = 3.44; C(4)=1.317.44≈5.429C(4) = 1.3\sqrt{17.44} \approx 5.429 (straight to FF under water). The cheapest route lands 0.50.5 km downstream from CC and then follows the bank for 3.53.5 km, for 3.443.44 million dollars. Compared with crossing straight to CC, the saving is 3.56−3.44=0.123.56 - 3.44 = 0.12 million dollars, that is 120 000120\,000 dollars.

d) Now C(x)=1.31.44+x2+0.5(0.4−x)C(x) = 1.3\sqrt{1.44 + x^2} + 0.5(0.4 - x) on [0,0.4][0, 0.4], with the same derivative. Its only zero, x=0.5x = 0.5, is outside the domain. Since x1.44+x2\frac{x}{\sqrt{1.44 + x^2}} increases with xx (its value is the sine of the angle ∠CSP\angle CSP, which grows with xx), C′(x)<C′(0.5)=0C'(x) < C'(0.5) = 0 on [0,0.4][0, 0.4]: CC is decreasing and its minimum is at the right endpoint x=0.4x = 0.4, the straight line from SS to the factory, entirely under water. Cost: 1.31.44+0.16=1.31.6≈1.64441.3\sqrt{1.44 + 0.16} = 1.3\sqrt{1.6} \approx 1.6444 million dollars, less than C(0)=1.56+0.2=1.76C(0) = 1.56 + 0.2 = 1.76. Answering x=0.5x = 0.5 here would lay pipe past the factory.

e) C′(x)=1.3xw2+x2−0.5=0C'(x) = \frac{1.3x}{\sqrt{w^2 + x^2}} - 0.5 = 0 gives, after isolating and squaring positive sides, 1.69x2=0.25(w2+x2)1.69x^2 = 0.25(w^2 + x^2), so 1.44x2=0.25w21.44x^2 = 0.25w^2 and x=0.5w1.2=5w12x = \frac{0.5w}{1.2} = \frac{5w}{12}. The distance to the factory does not appear: it only enters CC through −0.5x-0.5x plus a constant. For w=2.4w = 2.4 km, x=1x = 1 km; for w=1.2w = 1.2, x=0.5x = 0.5, as in b). The condition also reads xw2+x2=0.51.3=513\frac{x}{\sqrt{w^2 + x^2}} = \frac{0.5}{1.3} = \frac{5}{13}: the sine of the angle at SS equals the ratio of the two costs per km.

Exercise 8: Five statements to correct

Each statement below was written on a MATH 203 assignment, and each is false. Most of them are algebra, not calculus: that is where optimization marks are lost. Say what is wrong, give a counterexample, and write the correct statement.

  • a) To find the critical numbers of f(x)=x4−18x2f(x) = x^4 - 18x^2, solve 4x3=36x4x^3 = 36x by dividing both sides by 4x4x: x2=9x^2 = 9, so x=±3x = \pm 3.
  • b) 1.44+x2=1.2+x\sqrt{1.44 + x^2} = 1.2 + x, so the square root in the pipeline problem can be removed before differentiating.
  • c) The equation e−p/8(8−p)=0e^{-p/8}(8 - p) = 0 has two solutions: p=8p = 8, and the solution of e−p/8=0e^{-p/8} = 0.
  • d) If the objective function has exactly one critical number in its domain, that critical number gives its maximum.
  • e) Squaring both sides of an equation always gives an equivalent equation.

Type your answers, the page tells you right or wrong 0/5

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

Answers

  • a) False: 4x(x−3)(x+3)=04x(x - 3)(x + 3) = 0 has THREE solutions, −3-3, 00, 33.
  • b) False: at x=0.5x = 0.5, 1.69=1.3≠1.7\sqrt{1.69} = 1.3 \ne 1.7; a+b≠a+b\sqrt{a + b} \ne \sqrt a + \sqrt b.
  • c) False: e−p/8>0e^{-p/8} > 0 for every pp; the only solution is p=8p = 8.
  • d) False: the can of Exercise 3 has one critical number, a MINIMUM, and no maximum at all.
  • e) False: 1.3x=0.51.44+x21.3x = 0.5\sqrt{1.44 + x^2} has the root 0.50.5 only; its square also has −0.5-0.5.

a) FALSE. Dividing by 4x4x assumes x≠0x \ne 0 and deletes the solution x=0x = 0. Correct method: 4x3−36x=4x(x2−9)=4x(x−3)(x+3)=04x^3 - 36x = 4x(x^2 - 9) = 4x(x - 3)(x + 3) = 0, so the critical numbers are −3-3, 00 and 33. The lost one matters: f(0)=0f(0) = 0 is a local maximum of ff, and on [−1,1][-1, 1], say, it is the absolute maximum. Correct statement: bring everything to one side and FACTOR; never divide an equation by an expression that can be zero.

b) FALSE. At x=0.5x = 0.5: 1.44+0.25=1.69=1.3\sqrt{1.44 + 0.25} = \sqrt{1.69} = 1.3, while 1.2+0.5=1.71.2 + 0.5 = 1.7. The square root of a sum is not the sum of the square roots, just as (a+b)2≠a2+b2(a + b)^2 \ne a^2 + b^2. Correct statement: 1.44+x2\sqrt{1.44 + x^2} stays as it is, and is differentiated by the chain rule with inner function 1.44+x21.44 + x^2; the square root disappears only when the equation C′(x)=0C'(x) = 0 is squared, after the radical has been isolated.

c) FALSE. An exponential is never zero: eu>0e^{u} > 0 for every real uu, so e−p/8=0e^{-p/8} = 0 has NO solution. As p→∞p \to \infty, e−p/8→0e^{-p/8} \to 0 without ever reaching it, and a limit is not a solution. The equation has exactly one solution, p=8p = 8. Correct statement: in a product with an exponential factor, divide by that factor, which is legitimate precisely because it is never zero, and solve what is left. This is the opposite of a): dividing by e−p/8e^{-p/8} is safe, dividing by xx is not.

d) FALSE. For the open cup of Exercise 3, A(r)=πr2+2000rA(r) = \pi r^2 + \frac{2000}{r} has exactly one critical number on (0,∞)(0, \infty), and it gives the absolute MINIMUM; AA has no maximum, since A→∞A \to \infty at both ends of the domain. A single critical number can be a minimum, a maximum, or neither (x3x^3 at 00). Correct statement: if ff is continuous on an interval and has exactly one critical number cc there, and ff has a local maximum at cc (First Derivative Test), then f(c)f(c) is the absolute maximum on that interval.

e) FALSE. 1.3x=0.51.44+x21.3x = 0.5\sqrt{1.44 + x^2} has the single solution x=0.5x = 0.5; its square, 1.69x2=0.25(1.44+x2)1.69x^2 = 0.25(1.44 + x^2), has the solutions 0.50.5 and −0.5-0.5. For x=−0.5x = -0.5 the left side of the original is negative and the right side positive. Squaring loses the SIGN information, so it can create roots. Correct statement: squaring gives an equation whose solutions INCLUDE those of the original; every root of the squared equation must be checked in the original one, or both sides must be known to be ≥0\ge 0 on the domain.

Exercise 9: The price of a poke bowl: an exponential demand, factored

A food truck sells poke bowls. A season of sales data fits the demand model q(p)=600e−p/8q(p) = 600e^{-p/8}: at a price of pp dollars, it sells qq bowls per day. The figure shows the demand curve. The revenue is price times quantity, R=pqR = pq.

Every derivative in this exercise is a product with the factor e−p/8e^{-p/8}. Factoring it out, and remembering that it is never zero, is the gesture that makes each equation a one-liner. Money is given to the cent, with a calculator.

24681012141618202224100200300400500600700q = 600 e^(-p/8)(8, 600/e)p (dollars)q (bowls per day)
  • a) Write R(p)R(p), differentiate it with the product and chain rules, and factor: show that R′(p)=75e−p/8(8−p)R'(p) = 75e^{-p/8}(8 - p). Compute R′(0)R'(0) and interpret it.
  • b) Find the price that maximizes the daily revenue, justifying the absolute maximum on the domain [0,∞)[0, \infty), and give the maximal revenue to the cent.
  • c) Each bowl costs 33 dollars in ingredients, and the truck costs 400400 dollars per day to run. Find the price that maximizes the daily profit, and that profit to the cent.
  • d) Show that with an ingredient cost of cc dollars per bowl and any fixed daily cost, the best price is p=c+8p = c + 8. What is it for c=4.50c = 4.50 dollars?
  • e) The festival where the truck parks caps every price at 1010 dollars. Find the best price and the profit then, to the cent.

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

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

Answers

  • a) R(p)=600pe−p/8R(p) = 600pe^{-p/8}, R′(p)=600e−p/8(1−p8)=75e−p/8(8−p)R'(p) = 600e^{-p/8}\left(1 - \frac{p}{8}\right) = 75e^{-p/8}(8 - p); R′(0)=600R'(0) = 600
  • b) R′>0R' > 0 on [0,8)[0, 8), R′<0R' < 0 on (8,∞)(8, \infty): p=8p = 8 dollars, R=4800e≈1765.82R = \frac{4800}{e} \approx 1765.82 dollars
  • c) Π′(p)=75e−p/8(11−p)\Pi'(p) = 75e^{-p/8}(11 - p): p=11p = 11 dollars, Π=4800e−11/8−400≈813.63\Pi = 4800e^{-11/8} - 400 \approx 813.63 dollars
  • d) Π′(p)=75e−p/8(8+c−p)\Pi'(p) = 75e^{-p/8}(8 + c - p), the fixed cost disappears; c=4.50c = 4.50 gives p=12.50p = 12.50 dollars
  • e) Π′>0\Pi' > 0 on [0,10][0, 10]: p=10p = 10 dollars, Π=4200e−5/4−400≈803.32\Pi = 4200e^{-5/4} - 400 \approx 803.32 dollars

a) R(p)=p⋅600e−p/8=600pe−p/8R(p) = p \cdot 600e^{-p/8} = 600pe^{-p/8}. Product rule, and chain rule on e−p/8e^{-p/8} with inner function −p8-\frac{p}{8}, whose derivative is −18-\frac{1}{8}: R′(p)=600e−p/8+600p(−18)e−p/8R'(p) = 600e^{-p/8} + 600p\left(-\frac{1}{8}\right)e^{-p/8}. Factor the common factor 600e−p/8600e^{-p/8}: R′(p)=600e−p/8(1−p8)=75e−p/8(8−p)R'(p) = 600e^{-p/8}\left(1 - \frac{p}{8}\right) = 75e^{-p/8}(8 - p). R′(0)=75×8=600R'(0) = 75 \times 8 = 600: at a price near 00, each extra dollar on the price brings in about 600600 dollars more per day, one extra dollar on each of the 600600 bowls sold.

b) The factor 75e−p/875e^{-p/8} is positive for every pp, so the sign of R′R' is the sign of 8−p8 - p: positive on [0,8)[0, 8), negative on (8,∞)(8, \infty). RR increases then decreases on the WHOLE domain, so p=8p = 8 gives the absolute maximum; no endpoint at infinity needs to be evaluated. The truck sells 600e−1≈221600e^{-1} \approx 221 bowls per day, and the revenue is R(8)=4800e−1=4800e≈1765.82R(8) = 4800e^{-1} = \frac{4800}{e} \approx 1765.82 dollars. Writing the answer as p=8p = 8 alone would leave out the revenue, which was asked.

c) Π(p)=(p−3)⋅600e−p/8−400\Pi(p) = (p - 3) \cdot 600e^{-p/8} - 400. The fixed 400400 dollars differentiates to 00. Π′(p)=600e−p/8+(p−3)(−18)600e−p/8=75e−p/8[8−(p−3)]=75e−p/8(11−p)\Pi'(p) = 600e^{-p/8} + (p - 3)\left(-\frac{1}{8}\right)600e^{-p/8} = 75e^{-p/8}\left[8 - (p - 3)\right] = 75e^{-p/8}(11 - p). Watch the bracket: −(p−3)=−p+3-(p - 3) = -p + 3, and dropping the parentheses gives the false 5−p5 - p. The sign of Π′\Pi' is that of 11−p11 - p, positive then negative: p=11p = 11 dollars gives the absolute maximum. Then q=600e−11/8≈152q = 600e^{-11/8} \approx 152 bowls and Π(11)=8×600e−11/8−400=4800e−11/8−400≈813.63\Pi(11) = 8 \times 600e^{-11/8} - 400 = 4800e^{-11/8} - 400 \approx 813.63 dollars. The profit price is higher than the revenue price: each bowl now costs something, so selling fewer at a higher price pays.

d) Π(p)=(p−c)⋅600e−p/8−F\Pi(p) = (p - c) \cdot 600e^{-p/8} - F, so Π′(p)=75e−p/8[8−(p−c)]=75e−p/8(8+c−p)\Pi'(p) = 75e^{-p/8}\left[8 - (p - c)\right] = 75e^{-p/8}(8 + c - p): the fixed cost FF has vanished, and the only critical number, p=c+8p = c + 8, is again an absolute maximum by the sign of 8+c−p8 + c - p. Every dollar of ingredient cost raises the best price by exactly one dollar; the fixed cost changes the PROFIT, never the price. For c=4.50c = 4.50: p=12.50p = 12.50 dollars. With c=3c = 3, p=11p = 11, as in c).

e) The domain becomes [0,10][0, 10]. On it, 11−p≥1>011 - p \ge 1 > 0, so Π′>0\Pi' > 0 and Π\Pi is increasing: the best price is the cap itself, p=10p = 10 dollars, where Π′(10)≠0\Pi'(10) \ne 0. Π(10)=7×600e−5/4−400=4200e−5/4−400≈803.32\Pi(10) = 7 \times 600e^{-5/4} - 400 = 4200e^{-5/4} - 400 \approx 803.32 dollars, only about 1010 dollars less than without the cap, because Π\Pi is flat near its peak (solution figure). Solving Π′(p)=0\Pi'(p) = 0 and answering p=11p = 11 would break the festival rule.

24681012141618202224-400-2002004006008001000cap p = 10peak p = 11p (dollars)profit (dollars)

Exercise 10: A final exam problem: the largest cone inside a sphere

A right circular cone is inscribed in a sphere of radius 99 cm: its apex and the whole circle of its base lie on the sphere. The figure is the vertical cross-section through the axis: the circle of radius 99 centred at OO, and inside it the triangle of the cone, of height hh and base radius rr.

This is the shape of a long final exam question: the constraint comes from Pythagoras, the choice of the variable decides whether a square root appears, the Closed Interval Method gives the answer, then a restriction and a generalization test the method.

O9hrapex
  • a) Using the right triangle formed by OO, the centre of the base and a point of the base circle, show that r2=18h−h2r^2 = 18h - h^2, and deduce the volume V(h)V(h) of the cone and the domain of hh. Compute rr when h=3h = 3, to four decimals.
  • b) Find the cone of largest volume: its height, its base radius (to four decimals) and its volume, exactly and to the hundredth.
  • c) What fraction of the volume of the sphere does this cone fill?
  • d) The cone must fit in a display case: its height may not exceed 1010 cm. Find the largest volume now, to the hundredth.
  • e) For a sphere of radius RR, find the height of the largest inscribed cone, and show that the fraction of c) does not depend on RR.

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

a)
Domain of hh ,
b)
c)
d)
e)
Show the solution

Answers

  • a) r2=81−(h−9)2=18h−h2r^2 = 81 - (h - 9)^2 = 18h - h^2; V(h)=π3(18h2−h3)V(h) = \frac{\pi}{3}(18h^2 - h^3) on [0,18][0, 18]; r(3)=35≈6.7082r(3) = 3\sqrt 5 \approx 6.7082 cm
  • b) V′(h)=πh(12−h)V'(h) = \pi h(12 - h): h=12h = 12 cm, r=62≈8.4853r = 6\sqrt 2 \approx 8.4853 cm, V=288π≈904.78V = 288\pi \approx 904.78 cm3^3
  • c) 288π972π=827\frac{288\pi}{972\pi} = \frac{8}{27}
  • d) V′>0V' > 0 on (0,10](0, 10]: h=10h = 10 cm, V=800π3≈837.76V = \frac{800\pi}{3} \approx 837.76 cm3^3
  • e) V=π3(2Rh2−h3)V = \frac{\pi}{3}(2Rh^2 - h^3), h=4R3h = \frac{4R}{3}, V=32πR381=827⋅4πR33V = \frac{32\pi R^3}{81} = \frac{8}{27} \cdot \frac{4\pi R^3}{3}

a) Put the apex at the top of the sphere. The centre OO is 99 cm below the apex and the base plane hh cm below it, so OO is at distance ∣h−9∣|h - 9| from the centre of the base. Pythagoras in the right triangle OO, centre of the base, point of the base circle (hypotenuse 99): r2+(h−9)2=81r^2 + (h - 9)^2 = 81. Expand the square IN FULL, (h−9)2=h2−18h+81(h - 9)^2 = h^2 - 18h + 81, never h2−81h^2 - 81: r2=81−h2+18h−81=18h−h2r^2 = 81 - h^2 + 18h - 81 = 18h - h^2. The same formula holds whether the base is below or above OO, because the square hides the sign. Then V=13πr2h=π3(18h−h2)h=π3(18h2−h3)V = \frac{1}{3}\pi r^2 h = \frac{\pi}{3}(18h - h^2)h = \frac{\pi}{3}(18h^2 - h^3), on [0,18][0, 18] (at both ends the cone is flat: r=0r = 0). For h=3h = 3: r2=54−9=45r^2 = 54 - 9 = 45, r=35≈6.7082r = 3\sqrt 5 \approx 6.7082 cm. Check: h=0h = 0 gives r=0r = 0, as it must; the false expansion h2−81h^2 - 81 would give r2=162−h2r^2 = 162 - h^2 and r≈12.7r \approx 12.7 at h=0h = 0, wider than the sphere.

b) V′(h)=π3(36h−3h2)=πh(12−h)V'(h) = \frac{\pi}{3}(36h - 3h^2) = \pi h(12 - h). Factor the common factor hh, do not divide by it: the critical numbers are h=0h = 0 (an endpoint) and h=12h = 12. VV is a polynomial, continuous on the closed interval [0,18][0, 18], so the Closed Interval Method applies: V(0)=0V(0) = 0, V(18)=π3(5832−5832)=0V(18) = \frac{\pi}{3}(5832 - 5832) = 0, V(12)=π3(2592−1728)=864π3=288π≈904.78V(12) = \frac{\pi}{3}(2592 - 1728) = \frac{864\pi}{3} = 288\pi \approx 904.78 cm3^3. The largest cone has height 1212 cm and r2=216−144=72r^2 = 216 - 144 = 72, r=62≈8.4853r = 6\sqrt 2 \approx 8.4853 cm. Choosing hh as the variable is what kept the square root away: with rr as the variable, h=9+81−r2h = 9 + \sqrt{81 - r^2}, and the derivative needs the chain rule on a radical.

c) The sphere has volume 43π×93=972π\frac{4}{3}\pi \times 9^3 = 972\pi cm3^3, so the cone fills 288π972π=827≈0.296\frac{288\pi}{972\pi} = \frac{8}{27} \approx 0.296 of it, a little less than a third. The comparison is a sanity check too: an inscribed cone must fill less than the whole sphere.

d) The domain becomes [0,10][0, 10], and the critical number 1212 is outside it. On (0,10](0, 10], V′(h)=πh(12−h)>0V'(h) = \pi h(12 - h) > 0, since both factors are positive: VV is increasing and its maximum is at the endpoint h=10h = 10, where V′≠0V' \ne 0. There r2=180−100=80r^2 = 180 - 100 = 80 and V=π3(1800−1000)=800π3≈837.76V = \frac{\pi}{3}(1800 - 1000) = \frac{800\pi}{3} \approx 837.76 cm3^3. Answering h=12h = 12 would build a cone that does not fit in the case.

e) Same Pythagoras with radius RR: r2=R2−(h−R)2=2Rh−h2r^2 = R^2 - (h - R)^2 = 2Rh - h^2, so V=π3(2Rh2−h3)V = \frac{\pi}{3}(2Rh^2 - h^3) on [0,2R][0, 2R]. V′(h)=π3(4Rh−3h2)=π3h(4R−3h)V'(h) = \frac{\pi}{3}(4Rh - 3h^2) = \frac{\pi}{3}h(4R - 3h), zero at h=4R3h = \frac{4R}{3}, with V=0V = 0 at both endpoints: the absolute maximum. Then r2=8R23−16R29=8R29r^2 = \frac{8R^2}{3} - \frac{16R^2}{9} = \frac{8R^2}{9} and V=π3⋅8R29⋅4R3=32πR381V = \frac{\pi}{3} \cdot \frac{8R^2}{9} \cdot \frac{4R}{3} = \frac{32\pi R^3}{81}. Divided by 4πR33\frac{4\pi R^3}{3}: 3281⋅34=827\frac{32}{81} \cdot \frac{3}{4} = \frac{8}{27} for EVERY sphere. With R=9R = 9: h=12h = 12 and V=32π×72981=288πV = \frac{32\pi \times 729}{81} = 288\pi, as in b).

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