MATH 203 Calculus I • Concordia University, Montreal
Corrected exercises: limits involving infinity and asymptotes (MATH 203)
This is the corrected exercise set for limits involving infinity and asymptotes in MATH 203, Differential and Integral Calculus I, at Concordia University: section 2.6 of Thomas' Calculus, taught right after one-sided limits and before continuity. Limits at ±∞, infinite limits, horizontal, vertical and oblique asymptotes, dominant terms: every limit here is settled by algebra, never by L'Hôpital's Rule, which comes much later in the course. Your scientific calculator is useful to check a limit with a large value of x; it never replaces the argument.
The thread of the whole set: the limit rules of this chapter fit on five lines, and the points are lost in the algebra that has to happen before a rule applies. Dividing EVERY term by the dominant power, and subtracting exponents correctly when that power is x−1 or x2/3; factoring x2 out of a square root, where it comes out as ∣x∣; factoring the denominator (difference of squares, difference of cubes) before reading a sign; long division with a 0 placeholder and a subtraction that changes every sign; the laws of exponents and logarithms. Each solution names the gesture where it costs marks.
The traps named in the solutions: treating x−3 as dominant at infinity, writing x2=x at −∞, splitting x2+5 into x+5, forgetting that a negative numerator reverses both one-sided signs, declaring an asymptote where a factor cancels, simplifying exe2x into e2, reading 0⋅∞ or ∞−∞ as a value, believing lnx levels off, dividing only the first term of a numerator, and keeping the wrong root of a quadratic inequality.
Self-checking setType 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.
•limx→±∞xrc=0 for r>0. Quotient at ±∞: divide every term by the highest power of the DENOMINATOR; xbxa=xa−b.
•x2=∣x∣: for x<0, x2+⋯=−x1+⋯. But 3x3=x for every x.
•DN with N→L=0, D→0: infinite, with the sign of DL on EACH side. If N→0 too, factor and cancel first: a hole, not an asymptote.
•ex→0 at −∞, ex→∞ at ∞; lnx→−∞ at 0+, lnx→∞ at ∞. ebea=ea−b, lnA−lnB=lnBA.
•Numerator degree one more than the denominator: f(x)=mx+b+D(x)R(x), oblique asymptote y=mx+b; the sign of DR gives the side.
•∞−∞, ∞∞, 0⋅∞ are forms, not values: conjugate, factor, divide, combine logarithms, substitute t=x1, or sandwich.
Part A: the basics (/50)
Exercise 1: Limits at infinity: divide EVERY term by the dominant power
Thomas's basic facts: limx→±∞k=k and limx→±∞x1=0, so xrc→0 for every r>0 (at −∞ whenever xr is defined). For a quotient, divide the numerator AND the denominator, every term of each, by the highest power of x that appears in the DENOMINATOR, then let every xrc go to 0.
The same method works with negative and fractional exponents, once every term is written as a power of x: x=x1/2, x−3=x31, and xbxa=xa−b. The figure shows f(x)=x2+32x2−8.
a) Read on the figure limx→∞f(x) and limx→−∞f(x), then prove both. Does the graph of f ever reach its horizontal asymptote?
b) Find limx→∞2x2+x5−4x3 and limx→−∞2x2+x5−4x3.
c) Find limx→∞2x−2−x−16x−1+x−3.
d) Find limx→∞3x+x4x−x2/3+1.
e) Find limx→∞2x+9x3/2−7x and limx→∞x3/2+23x+1.
Show the solution
Answers
a)2 at both ends; f(x)−2=−x2+314<0, so the graph never reaches y=2 and approaches it from below.
b)−∞ as x→∞, +∞ as x→−∞: no horizontal asymptote.
c)−6
d)34
e)+∞ and 0
a) The figure shows the curve flattening onto the line y=2 at both ends. Proof: the highest power of the denominator is x2, and dividing EVERY term by it gives f(x)=1+x232−x28→1+02−0=2 as x→∞ and as x→−∞. So y=2 is the horizontal asymptote at both ends. To see whether the graph reaches it, compute the gap on a common denominator: f(x)−2=x2+32x2−8−2(x2+3)=x2+32x2−8−2x2−6=−x2+314. The −2 must multiply BOTH terms of x2+3: writing −2x2+6 gives a gap of −2 and a wrong conclusion. The gap is negative for every x, so the graph stays strictly below y=2 and never reaches it.
b) Divide every term by x2: 2x2+x5−4x3=2+x1x25−4x. The denominator tends to 2. As x→∞, −4x→−∞, so the quotient tends to −∞; as x→−∞, −4x→+∞, so the quotient tends to +∞. The shortcut, and the check, is the quotient of the dominant terms: 2x2−4x3=−2x, which is odd, so the two ends go opposite ways. Degree 3 over degree 2: no horizontal asymptote.
c) Rewrite the exponents before choosing anything: 2x−2−x−16x−1+x−3=x22−x1x6+x31. Clear the small fractions by multiplying the numerator and the denominator by x3: 2x−x26x2+1. Now divide by x2: x2−16+x21→−16=−6. The trap is to treat x−3 as the dominant term because 3>1: at infinity, x−3=x31 is the SMALLEST term, and x−1 the largest. Pairing the coefficients as they are written, 26=3, is the other classic loss.
d) As x→∞, x is larger than x=x1/2, so the dominant power of the denominator is x. Divide every term by x, subtracting exponents: xx2/3=x2/3−1=x−1/3=3x1 and xx=x−1/2=x1. Then 3+x−1/24−x−1/3+x1→34. The exponent rule is xbxa=xa−b, not xa/b: xx2/3 is not x2/3.
e) First limit: the dominant power of the denominator is x; dividing, 2+x9x1/2−7, whose numerator tends to ∞ because x→∞. The limit is +∞: the numerator's largest exponent, 23, beats the denominator's, 1. Second limit: now the denominator's dominant power is x3/2: 1+2x−3/23x−1/2+x−3/2→10=0. The rule of degrees for rational functions survives with fractional exponents: compare the largest exponent on top with the largest exponent below.
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: Square roots at infinity: the root of x squared is |x|
To divide a square root by a power of x, factor x2 out of the root: 9x2+4=x2(9+x24)=x29+x24=∣x∣9+x24. For x>0, ∣x∣=x; for x<0, ∣x∣=−x, and a minus sign appears that decides the limit at −∞.
A cube root behaves differently: 3x3=x for every real x, with no absolute value. The figure shows g(x)=2x−59x2+4.
a) Find limx→∞g(x).
b) Find limx→−∞g(x). How many horizontal asymptotes does the graph of g have?
c) Find limx→(5/2)−g(x) and limx→(5/2)+g(x). Can the graph of g have a hole instead?
d) Find limx→∞(x2+3x−x2−x) and limx→−∞(x2+3x−x2−x).
e) Find limx→∞x+138x3−x2 and limx→−∞x+138x3−x2.
Show the solution
Answers
a)23
b)−23; two horizontal asymptotes, y=23 on the right and y=−23 on the left.
c)−∞ and +∞; no hole is possible, since 9x2+4≥4 never vanishes.
d)2 and −2
e)2 at both ends: one horizontal asymptote, y=2.
a) For x>0, 9x2+4=x9+x24, and 2x−5=x(2−x5). Cancel x: g(x)=2−x59+x24→29=23. The root of the limit is the limit of the root, since is continuous at 9.
b) For x<0, x2=∣x∣=−x, so 9x2+4=−x9+x24. Then g(x)=x(2−x5)−x9+x24=−2−x59+x24→−23. The graph has TWO horizontal asymptotes, y=23 as x→∞ and y=−23 as x→−∞, as the figure shows. Sanity check before any algebra: at x=−100 the root is positive and 2x−5 is negative, so g<0, and an answer of +23 is impossible. Writing x2=x is the algebra slip that costs this whole limit.
c) As x→25, the numerator tends to 9⋅425+4=4241≈7.76>0, and 2x−5→0: through negative values on the left (0−), positive on the right (0+). So limx→(5/2)−g(x)=−∞ and limx→(5/2)+g(x)=+∞, and x=25 is a vertical asymptote. A hole would need the numerator to tend to 0 too, and 9x2+4≥4 for every x: impossible.
d) Both roots tend to +∞ at each end: the form ∞−∞. Multiply and divide by the conjugate: x2+3x−x2−x=x2+3x+x2−x(x2+3x)−(x2−x)=x2+3x+x2−x4x. The parentheses matter: subtracting the WHOLE of x2−x gives +x, hence 4x, not 2x. For x>0, each root is x⋯: 1+x3+1−x14→24=2. For x<0, each root is −x⋯, so the denominator is −x(1+x3+1−x1) and the quotient is 1+x3+1−x1−4→−2. Calculator check at x=100: 10300−9900≈101.489−99.499=1.990.
e) 38x3−x2=3x3(8−x1)=x38−x1, and this holds for EVERY x=0, negative or positive, because 3x3=x. With x+1=x(1+x1), the quotient is 1+x138−x1→38=2 at both ends. One horizontal asymptote, y=2. Contrast with b): an EVEN root turns x into ∣x∣ and can split the two ends; an ODD root keeps the sign of x and does not.
Exercise 3: Infinite limits: factor the denominator, then read the sign on each side
The line x=a is a vertical asymptote of y=f(x) when limx→a+f(x)=±∞ or limx→a−f(x)=±∞. For a quotient whose numerator tends to L=0 and whose denominator tends to 0, the size is infinite and the SIGN on each side is the whole question.
To read those signs, factor the denominator completely first: difference of squares, trinomial, difference of cubes a3−b3=(a−b)(a2+ab+b2). When the numerator ALSO tends to 0, factor it too and cancel before deciding. The figure shows f(x)=x2−4xx+1.
a) Read on the figure the one-sided limits of f at 0 and at 4, and limx→±∞f(x). Confirm the four one-sided limits from the factored formula.
b) Find limx→2−x2−43−2x and limx→2+x2−43−2x.
c) Let k(x)=x2−4x3−8. For x=2 and x=−2, decide whether the graph of k has a vertical asymptote there, and give the limits.
d) Find limx→0x2/31, limx→0−x1/31 and limx→0+x1/31.
e) Find limx→1−lnx1, limx→1+lnx1, limx→π−cotx and limx→π+cotx.
Show the solution
Answers
a)+∞ as x→0−, −∞ as x→0+; −∞ as x→4−, +∞ as x→4+; 0 at both ends.
b)+∞ from the left, −∞ from the right.
c)x=2: a hole, limx→2k(x)=3; x=−2: asymptote, −∞ on the left and +∞ on the right.
d)+∞; −∞ and +∞.
e)−∞, +∞; −∞, +∞. Asymptotes x=1 for lnx1 and x=π for cotx.
a) On the figure: along x=0 the curve climbs on the left and plunges on the right; along x=4 it plunges on the left and climbs on the right; it flattens onto the x-axis at both ends. From the formula, factor first: f(x)=x(x−4)x+1. Near 0 the numerator tends to 1>0 and x−4→−4<0. For x<0, x(x−4) is (negative)(negative), so the denominator tends to 0+ and f→+∞; for 0<x<4 it is (positive)(negative), 0−, and f→−∞. Near 4 the numerator tends to 5 and x→4, so the sign is that of x−4: 0− on the left, f→−∞; 0+ on the right, f→+∞. At ±∞ the degrees are 1 over 2, so f→0. Without the factorization, the sign of x2−4x near 0 is a guess.
b) The numerator tends to 3−4=−1. Factor the denominator: x2−4=(x−2)(x+2), with x+2→4. As x→2−, x−2→0−, so the denominator tends to 0− and the quotient is (negative)/(small negative): +∞. As x→2+ the denominator tends to 0+: −∞. A negative numerator reverses both signs; checking only the denominator gives the two answers backwards.
c) Factor both: x3−8=(x−2)(x2+2x+4) (difference of cubes, check by expanding) and x2−4=(x−2)(x+2). For x=2, k(x)=x+2x2+2x+4. At x=2 the original formula is a form 00, and after cancelling limx→2k(x)=44+4+4=3: a hole at (2,3), no asymptote. At x=−2 the simplified numerator tends to 4−4+4=4>0 and x+2→0− on the left, 0+ on the right: −∞ then +∞, and x=−2 is a vertical asymptote. The factorizations (x−2)3 or (x−2)(x2+4) are the usual slips: expand back to catch them.
d) x2/3=(3x)2 is positive on both sides of 0 and tends to 0, so x2/31→+∞ from both sides and limx→0x2/31=+∞. But x1/3=3x has the sign of x: it tends to 0− on the left and 0+ on the right, so x1/31 tends to −∞ on the left and +∞ on the right, and limx→0x1/31 does not exist. Reading a fractional exponent means reading its root: an odd root keeps the sign, an even power removes it.
e) ln1=0, lnx<0 for 0<x<1 and lnx>0 for x>1. So lnx1→0−1=−∞ as x→1− and +∞ as x→1+: x=1 is a vertical asymptote. Next, cotx=sinxcosx with cosx→−1 as x→π; sinx>0 just left of π and <0 just right. Hence cotx→0+−1=−∞ as x→π− and 0−−1=+∞ as x→π+: x=π is a vertical asymptote. The sign of sin near π comes from the unit circle, not from the calculator.
Exercise 4: Exponentials and logarithms at the ends of the axis
The facts read on the graphs: ex→∞ as x→∞ and ex→0 as x→−∞, so e−x does the opposite; lnx→∞ as x→∞ and lnx→−∞ as x→0+. The algebra that goes with them: e2x=(ex)2, ebea=ea−b, e−x=ex1, lnA−lnB=lnBA and klnx=lnxk for x>0.
Two tools of Thomas 2.6 also appear: the Sandwich Theorem, which works as x→±∞ exactly as at a point, and the substitution t=x1, which turns x→∞ into t→0+. The figure shows h(x)=1+ex5−ex.
a) Find limx→∞h(x) and limx→−∞h(x). Find exactly where the graph of h crosses the x-axis.
b) Find limx→∞e2x−exe2x+1 and limx→−∞e2x−exe2x+1.
c) Using t=x1, find limx→∞e1/x and limx→∞3xtan(x1).
d) Find limx→∞[2lnx−ln(4x2+1)] and limx→0+[2lnx−ln(4x2+1)].
e) Find limx→∞e−xcosx. What happens to e−xcosx as x→−∞?
Show the solution
Answers
a)−1 and 5; the graph crosses the x-axis at x=ln5≈1.6094.
b)1 and −∞
c)1 and 3
d)ln41=−ln4≈−1.3863, and −∞ (not an indeterminate form).
e)0 by the Sandwich Theorem; as x→−∞ it oscillates with growing amplitude: no limit, not even an infinite one.
a) As x→∞, ex dominates: divide every term by it. h(x)=e−x+15e−x−1→0+10−1=−1. As x→−∞, ex→0 and nothing needs dividing: h(x)→1+05−0=5. Two horizontal asymptotes, y=5 on the left and y=−1 on the right. The crossing: h(x)=0 exactly when 5−ex=0, since 1+ex>0; so ex=5 and x=ln5≈1.6094. Note that e−x+15e−x−1 comes from ex5=5e−x and exex=1: each term divided, each exponent subtracted.
b) As x→∞ the dominant exponential is e2x: 1−e−x1+e−2x→1, because e2xex=ex−2x=e−x. As x→−∞, the numerator tends to 0+1=1 and the denominator to 0−0=0: study its sign. Factor: e2x−ex=ex(ex−1), with ex>0 and ex−1<0 for x<0, so the denominator tends to 0− and the quotient to −∞. Dividing by e2x at −∞ would be dividing by something that tends to 0: the dominant term depends on the direction.
c) With t=x1, x→∞ means t→0+. First, e1/x=et→e0=1, by continuity of the exponential. Second, 3xtan(x1)=t3tant=3⋅tsint⋅cost1→3⋅1⋅1=3, using limt→0tsint=1 from section 2.4. As written, the limit is the form ∞⋅0 (x→∞, tanx1→0), which decides nothing; the substitution turns it into a limit we know.
d) As x→∞ both logarithms tend to ∞: the form ∞−∞. Use the laws of logarithms, valid for x>0: 2lnx=lnx2 and lnx2−ln(4x2+1)=ln4x2+1x2. The inside tends to 41 (equal degrees), and ln is continuous at 41, so the limit is ln41=−ln4≈−1.3863. The false law ln(4x2+1)=ln4x2+ln1 is the slip to avoid: the logarithm of a SUM does not split. As x→0+, 2lnx→−∞ while ln(4x2+1)→ln1=0: the difference tends to −∞, and there is nothing indeterminate to transform.
e) For every x, −1≤cosx≤1, and e−x>0, so −e−x≤e−xcosx≤e−x. Both bounds tend to 0 as x→∞, so by the Sandwich Theorem limx→∞e−xcosx=0: y=0 is a horizontal asymptote, crossed at every zero of cosx. As x→−∞, e−x→∞: at x=−2kπ the function equals e2kπ, which grows without bound, and at x=−(2k+1)π it equals −e(2k+1)π, which goes to −∞. The values are neither close to one number nor all large and positive: the limit does not exist, not even as ∞.
Exercise 5: Oblique asymptotes and dominant terms: long division done right
When the degree of the numerator of a rational function is exactly ONE more than the degree of the denominator, the graph has an oblique (slant) asymptote. Long division gives f(x)=(mx+b)+D(x)R(x) with degR<degD; the fraction tends to 0 at both ends, and Thomas calls mx+b the dominant term of f for large ∣x∣. Near a vertical asymptote, the dominant term is the one that blows up.
Two algebra rules make or break the division: a missing power is written with a 0 placeholder (4x3+0x2−2x+5), and subtracting a line changes the sign of EVERY term of that line. The figure shows f(x)=x−22x2−3x−1 and the dashed line y=2x+1.
a) Divide 2x2−3x−1 by x−2 and deduce the oblique asymptote of f. Give its vertical asymptote with the two one-sided limits.
b) On which side of its oblique asymptote is the graph of f, for x>2 and for x<2?
c) Let q(x)=2x2−x4x3−2x+5. Find its oblique asymptote and the point where the graph crosses it. How many vertical asymptotes does q have?
d) Let s(x)=x2x3−4x+6. Write s(x) as a sum of terms, then give its dominant term for large ∣x∣ and its behaviour near x=0.
e) Let u(x)=2x−3+e−x. Is the line y=2x−3 an asymptote of u as x→∞? As x→−∞?
Show the solution
Answers
a)f(x)=2x+1+x−21: oblique asymptote y=2x+1; vertical asymptote x=2, −∞ on the left, +∞ on the right.
b)Above for x>2, below for x<2, since f(x)−(2x+1)=x−21.
c)q(x)=2x+1+2x2−x5−x: asymptote y=2x+1, crossed at (5,11); two vertical asymptotes, x=0 and x=21.
d)s(x)=x−x4+x26: dominant term x (asymptote y=x); s(x)→+∞ on both sides of 0.
e)Yes as x→∞ (the gap e−x→0); no as x→−∞ (the gap e−x→∞).
a) Long division: 2x2÷x=2x, and 2x2−3x−1−2x(x−2)=2x2−3x−1−2x2+4x=x−1; then x÷x=1, and x−1−1⋅(x−2)=1. So 2x2−3x−1=(x−2)(2x+1)+1 and f(x)=2x+1+x−21. Check at x=3: f(3)=118−9−1=8 and 6+1+1=8. As x→±∞, x−21→0, so y=2x+1 is the oblique asymptote at both ends. At x=2 the numerator is 8−6−1=1=0: with x−2→0− on the left, f→−∞; with 0+ on the right, f→+∞. The subtraction step is where the division goes wrong: −2x(x−2)=−2x2+4x, and forgetting to change the sign of −4x gives the wrong asymptote y=2x−7.
b) f(x)−(2x+1)=x−21, positive for x>2 and negative for x<2. So the right branch lies ABOVE the dashed line and comes down onto it, and the left branch lies BELOW it and climbs to it, exactly as on the figure. The sign of the remainder term is what makes a sketch near an oblique asymptote correct.
c) Write the missing power: 4x3+0x2−2x+5. First step: 4x3÷2x2=2x, and 2x(2x2−x)=4x3−2x2; subtracting leaves 0x2+2x2−2x+5=2x2−2x+5. Second step: 2x2÷2x2=1, and subtracting 2x2−x leaves −x+5. So q(x)=2x+1+2x2−x5−x, and since the fraction tends to 0 (degree 1 over 2), y=2x+1 is the oblique asymptote. The gap vanishes when 5−x=0: at x=5, q(5)=50−5500−10+5=45495=11=2⋅5+1, so the graph crosses its asymptote at (5,11). The denominator x(2x−1) vanishes at 0 and 21, where the numerator equals 5 and 21−1+5=4.5, both nonzero: two vertical asymptotes. Without the 0x2 placeholder, the −2x2 of the first subtraction lands under −2x and the division collapses.
d) The denominator is a single term, so divide EACH term of the numerator by it: s(x)=x2x3−x24x+x26=x−x4+x26. For large ∣x∣ the last two terms tend to 0: the dominant term is x, and y=x is the oblique asymptote at both ends. Near x=0 the dominant term is x26, positive on both sides, so s(x)→+∞ as x→0− and as x→0+ (directly: the numerator tends to 6 and x2→0+). The slip to avoid is dividing only the first term: x2x3−4x+6=x−4x+6 is false.
e) u(x)−(2x−3)=e−x. As x→∞, e−x→0, so y=2x−3 is an oblique asymptote on the right, approached from above since e−x>0. As x→−∞, e−x→∞: the gap grows without bound, so the line is NOT an asymptote on the left. An oblique asymptote, like a horizontal one, belongs to ONE direction at a time; rational functions hide this because their two ends always agree.
Part B: problems and reasoning (/50)
Exercise 6: Asymptotes as equations: finding the constants
A final-exam favourite turns the chapter around: the asymptotes are given, the formula carries unknown constants, and each asymptote becomes an equation. Every candidate vertical asymptote must then be checked in the numerator, since a common zero gives a hole instead.
No figure: every answer comes from the formula.
a) Find a and b so that the graph of f(x)=x2+bxax2−3 has the horizontal asymptote y=−2 and the vertical asymptote x=4. Does f have another vertical asymptote?
b) Let h(x)=x2−5x+6x−k. For which values of k does the graph of h have exactly ONE vertical asymptote? For the smaller such k, describe the graph near its other excluded point.
c) Find a and b>0 so that the graph of g(x)=ex+baex+6 has the horizontal asymptote y=4 as x→∞ and y=3 as x→−∞. How many vertical asymptotes does g have?
d) Find k so that the graph of p(x)=4x2+3kx+1 has the horizontal asymptote y=5 as x→−∞. What is then its horizontal asymptote as x→∞?
Show the solution
Answers
a)a=−2, b=−4: f(x)=x(x−4)−2x2−3, with a second vertical asymptote x=0.
b)k=2 or k=3. For k=2: h(x)=x−31 for x=2, a hole at (2,−1) and the asymptote x=3.
c)a=4, b=2; no vertical asymptote, since ex+2>0.
d)k=−10; then y=−5 as x→∞.
a) Equal degrees, so f(x)→1a=a at both ends: a=−2. A vertical asymptote at 4 needs the denominator to vanish there: 16+4b=0, so b=−4 and the denominator is x2−4x=x(x−4). The numerator −2x2−3 is negative for every x, so it vanishes neither at 4 nor at 0: both x=4 and x=0 are vertical asymptotes, and there is no hole. The factorization of the denominator is what reveals the second asymptote, which the question does not announce.
b) Factor: x2−5x+6=(x−2)(x−3), so the candidates are x=2 and x=3. For a generic k the numerator x−k vanishes at neither, and there are two vertical asymptotes. If k=2, the factor cancels: h(x)=(x−2)(x−3)x−2=x−31 for x=2, which has only the asymptote x=3, and at x=2 a hole at (2,2−31)=(2,−1). If k=3, likewise h(x)=x−21 for x=3, a hole at (3,1) and the asymptote x=2. So exactly one vertical asymptote for k=2 and for k=3, and for no other k.
c) As x→∞, divide every term by ex: 1+be−xa+6e−x→a, so a=4. As x→−∞, ex→0: g(x)→0+b0+6=b6, so b6=3 and b=2. Then ex+2>2 never vanishes, so g has no vertical asymptote. Had the conditions forced b=−2, the denominator would vanish at x=ln2 and create one: the sign of b matters.
d) As x→−∞, 4x2+3=∣x∣4+x23=−x4+x23, so p(x)=−x4+x23x(k+x1)=−4+x23k+x1→−2k. Setting −2k=5 gives k=−10. As x→∞ the minus sign disappears and p(x)→2k=−5. A student who forgets ∣x∣=−x finds k=10 and swaps the two asymptotes: check with x=−100, where p(−100)=400031001>0, consistent with y=5 on the left.
Exercise 7: Creating graphs from limits, and limits from graphs
The figure shows the whole graph of a function f defined on (−∞,−1)∪(−1,∞): dashed lines mark x=−1, x=3, y=2 and y=−1, and the dot is the point of the graph at x=3.
Thomas's exercises on creating graphs work in both directions: read the limits on a graph, and choose a formula that produces given limits. A formula is checked limit by limit, never at a glance.
a) Read on the graph limx→−∞f(x), limx→∞f(x), limx→−1−f(x), limx→−1+f(x) and limx→3f(x).
b) Give f(3). Is x=3 a vertical asymptote, although f is defined there? How many asymptotes does the graph have in all?
c) Which function satisfies ALL of: the only vertical asymptote is x=1, with limx→1±=+∞; limx→±∞=3; and the value 0 at x=0? Choose among (x−1)23x2, (x−1)23x, x2−13x2, x23(x−1)2 and (x−1)23x2+1.
d) Which function has limx→±∞=0, limx→2−=−∞, limx→2+=+∞, limx→−2−=+∞ and limx→−2+=−∞? Choose among x2−41, x2−4x, x2−4x2 and (x−2)2(x+2)1.
e) Could the function of the figure be a rational function (a quotient of two polynomials)? Explain.
Show the solution
Answers
a)2; −1; +∞; −∞; +∞ (from both sides).
b)f(3)=1; yes, x=3 is a vertical asymptote; four asymptotes: x=−1, x=3, y=2, y=−1.
c)(x−1)23x2
d)x2−41
e)No: a rational function has the same limit at ∞ and at −∞ (or none), and this graph has two different horizontal asymptotes.
a) Far to the left the curve settles on the dashed line y=2, so limx→−∞f(x)=2; far to the right it settles on y=−1, so limx→∞f(x)=−1. Along x=−1 the left branch climbs out of the frame and the middle branch comes up from the bottom: limx→−1−f(x)=+∞ and limx→−1+f(x)=−∞. Along x=3 both neighbouring branches climb out of the frame: limx→3−f(x)=limx→3+f(x)=+∞, so limx→3f(x)=+∞.
b) The dot gives f(3)=1. The line x=3 is nevertheless a vertical asymptote: the definition asks about the one-sided LIMITS at 3, which are infinite, and never about the value at 3. The graph has four asymptotes: the vertical lines x=−1 and x=3, and the horizontal lines y=2 (on the left) and y=−1 (on the right).
c) Test each condition in turn. (x−1)23x2: equal degrees with leading coefficients 3 and 1, so the limit is 3 at both ends; near 1 the numerator tends to 3>0 and (x−1)2→0+ on both sides, so +∞ on both sides; the only zero of the denominator is 1; and its value at 0 is 0. It satisfies everything. The others fail: (x−1)23x tends to 0 at infinity (degree 1 over 2); x2−13x2 has a second asymptote x=−1 and changes sign at 1; x23(x−1)2 has its asymptote at x=0, where it is not even defined; (x−1)23x2+1 equals 1 at x=0.
d) x2−41=(x−2)(x+2)1 tends to 0 at both ends. Near 2, x+2→4 and x−2→0∓: −∞ on the left, +∞ on the right. Near −2, x−2→−4 and x+2→0− on the left, so the product tends to 0+ and the quotient to +∞; on the right the product tends to 0− and the quotient to −∞. All five conditions hold. x2−4x fails at −2: its numerator tends to −2<0, which reverses both signs there. x2−4x2 tends to 1 at infinity. (x−2)2(x+2)1 tends to +∞ on BOTH sides of 2, because of the square.
e) No. For a rational function QP, dividing by the highest power of the denominator gives the same result as x→∞ and as x→−∞ when the limit is finite: 0, or the ratio of the leading coefficients. So a rational function has at most ONE horizontal asymptote, shared by both ends. The graph of the figure approaches y=2 on the left and y=−1 on the right, so it is not the graph of a rational function: a root, an exponential or a piecewise definition is needed (here the three branches come from three different formulas).
Exercise 8: Five statements to correct
Each statement below was written on a MATH 203 assignment, and each is false. Say what is wrong, give a counterexample or the correct computation, and write the correct statement. Four of the five errors are algebra, not calculus.
a) x2+5=x+5 for x>0, so limx→∞(x2+5−x)=5.
b) The graph of x2−9x−3 has two vertical asymptotes, x=3 and x=−3.
c) limx→∞ex+1e2x=e2, because exe2x=e2.
d) If limx→∞f(x)=0 and limx→∞g(x)=∞, then limx→∞f(x)g(x)=0.
e) The graph of y=lnx flattens out, so it has a horizontal asymptote as x→∞.
Show the solution
Answers
a)False: a+b=a+b; by the conjugate, x2+5−x=x2+5+x5→0.
b)False: x2−9x−3=x+31 for x=3; a hole at (3,61) and only one asymptote, x=−3.
c)False: exe2x=ex; the quotient equals 1+e−xex→∞.
d)False: 0⋅∞ is indeterminate; x1⋅x2→∞, x3⋅x→3.
e)False: lnx>M as soon as x>eM, so lnx→∞: no horizontal asymptote.
a) FALSE. A square root does not split over a sum: at x=2, 4+5=3 while 2+5≈4.24. The limit is a form ∞−∞ to settle by the conjugate: x2+5−x=x2+5+x(x2+5)−x2=x2+5+x5, whose denominator tends to ∞, so the limit is 0. Correct statement: x2+5−x→0 as x→∞; in general a+b=a+b.
b) FALSE. Factor before deciding: x2−9=(x−3)(x+3), so x2−9x−3=x+31 for x=3 and the limit at 3 is 61: the graph has a hole at (3,61), not an asymptote. At −3 the simplified numerator is 1=0, so x=−3 is the only vertical asymptote. Correct statement: a zero of the denominator gives a vertical asymptote only if the numerator does not tend to 0 there.
c) FALSE. The exponent law is ebea=ea−b, so exe2x=e2x−x=ex, not e2 (that would be dividing the exponents and dropping x). Dividing every term by ex: ex+1e2x=1+e−xex, whose numerator tends to ∞ and denominator to 1. Correct statement: limx→∞ex+1e2x=∞, no horizontal asymptote on the right.
d) FALSE. 0⋅∞ is an indeterminate form: the answer depends on how fast each factor goes. With f(x)=x1 and g(x)=x2, fg=x→∞; with f(x)=x3 and g(x)=x, fg=3; with f(x)=x21 and g(x)=x, fg=x1→0. Correct statement: the product must be simplified into one expression before its limit is taken.
e) FALSE. lnx grows slowly but WITHOUT BOUND: given any level M, lnx>M for every x>eM, since ln is increasing. For instance lnx>100 for x>e100. So limx→∞lnx=∞ and no horizontal line is approached. The flattening seen on a calculator window is an effect of the window. Correct statement: lnx→∞ as x→∞; the graph of lnx has only the vertical asymptote x=0.
Exercise 9: Average cost: what happens when production grows
A workshop assembles electric bicycles. Its fixed costs are 18000 dollars a month (rent, salaries, equipment), and the parts of each bicycle cost 240 dollars. In a first model, producing x bicycles in a month costs C1(x)=18000+240x dollars. The AVERAGE COST per bicycle is A(x)=xC(x).
Above a certain volume the workshop pays overtime and rents extra space; a second model adds a term for that: C2(x)=18000+240x+0.05x2. A third option is an automated line, with C3(x)=60000+200x+3000x. The figure shows the average costs A1 and A2 of the first two models. Give amounts to the cent when they are not whole.
a) Write A1(x) as a sum of simple terms. Find limx→∞A1(x) and limx→0+A1(x), and interpret both.
b) Show that A1(x)>240 for every x>0. What is the smallest monthly production for which the average cost is below 250 dollars?
c) Write A2(x) as a sum of simple terms and show that its graph has an oblique asymptote. Give it.
d) Find limx→∞A2(x). For which productions is A2(x) within 5 dollars of its oblique asymptote?
e) Find limx→∞A3(x) and A3(500). Which option has the lower average cost in the long run, the automated line or model 1?
Show the solution
Answers
a)A1(x)=x18000+240; limx→∞A1=240 dollars (the cost of the parts), limx→0+A1=+∞.
b)A1(x)−240=x18000>0; below 250 dollars from x=1801 bicycles on.
e)200 dollars; A3(500)≈454.16 dollars; the automated line, since 200<240.
a) The denominator x is a single term, so divide EACH term of the cost by it: A1(x)=x18000+240x=x18000+240. As x→∞, x18000→0, so A1(x)→240: the line y=240 is a horizontal asymptote. In words, the fixed costs are spread over more and more bicycles and their share per bicycle vanishes; what remains is the 240 dollars of parts that every bicycle needs. As x→0+, x18000→+∞: the y-axis is a vertical asymptote, and a workshop that builds almost nothing pays an enormous amount per bicycle. Writing x18000+240x=18000+240 (cancelling x in one term only) is the algebra slip this part is testing.
b) A1(x)−240=x18000, positive for every x>0: the average cost always stays above its asymptote and approaches it from above, never reaching it. Next, A1(x)<250 means x18000<10, that is x>1800 (multiplying by x>0 keeps the direction of the inequality). At exactly 1800 bicycles the average cost is 250 dollars, not below, so the smallest production is 1801 bicycles, where A1=180118000+240≈249.99 dollars.
c) Dividing each term by x: A2(x)=x18000+240+0.05x. The gap between A2 and the line y=0.05x+240 is A2(x)−(0.05x+240)=x18000, which tends to 0 as x→∞. So y=0.05x+240 is an oblique asymptote, approached from above. Here no long division is needed: when the denominator is a single power of x, splitting the fraction term by term IS the division.
d) A2(x)≥240+0.05x, which tends to ∞, so limx→∞A2(x)=+∞ and there is no horizontal asymptote: for large volumes the overtime term 0.05x dominates and the average cost climbs again, as the figure shows. Within 5 dollars of the asymptote means x18000<5, that is x>3600 bicycles. The dominant terms tell the whole story: x18000 near 0, 0.05x for large x.
e) Divide each term by x, with xx=x1/2−1=x−1/2=x1: A3(x)=x60000+200+x3000. Both fractions tend to 0, so limx→∞A3(x)=200 dollars. At x=500: A3(500)=120+200+5003000≈120+200+134.16=454.16 dollars, while A1(500)=36+240=276 dollars. So the automated line is much more expensive at 500 bicycles, but its asymptote 200 is below the 240 of model 1: in the long run it wins. At 10000 bicycles, A3=6+200+30=236 dollars against A1=241.80 dollars. The limits compare the two options for very large volumes only; the values at a realistic x are what decide a purchase.
Exercise 10: A drug in the long term: washout and steady state
The same drug can be given in two ways. After a single injection, the concentration in the blood is modelled by CA(t)=t2+1640t mg/L, t hours after the injection. A skin patch releases the drug continuously, and the concentration is then modelled by CB(t)=t2+2520t mg/L, t hours after the patch is applied.
The questions are about the long term only: what each concentration tends to, and how long that takes. Round times to the hundredth of an hour. The figure shows both models over two days.
a) Find limt→∞CA(t) and interpret it. What is the dominant term of CA(t) for large t?
b) From what time on does the concentration after the injection stay below 2 mg/L?
c) Find limt→∞CB(t), the steady-state concentration of the patch. Show that it is never reached, and find when CB(t)=16 mg/L.
d) From what time on is CB(t) within 0.5 mg/L of its steady state?
e) A patient receives the injection and the patch at the same moment, so that C(t)=CA(t)+CB(t). Find limt→∞C(t) and limt→∞CB(t)CA(t). Does C(t) approach its limit from above or from below?
Show the solution
Answers
a)0 mg/L: the drug is eliminated; dominant term t40.
b)From t=10+221≈19.17 h on (the other root, 0.83 h, is during the rise).
c)20 mg/L; 20−CB(t)>0 since t<t2+25; CB=16 at t=320≈6.67 h.
d)From t≈21.94 h on.
e)20 mg/L and 0; from above, since CA≈t40 outweighs 20−CB≈t2250.
a) Divide every term by t2, the dominant power of the denominator: CA(t)=1+t216t40→10=0. The line C=0 is a horizontal asymptote: in the long run the drug is eliminated (washout). For large t, t2+16 is dominated by t2, so CA(t) behaves like t240t=t40: at t=100, CA=100164000≈0.3994 against 10040=0.4. The dominant term says HOW FAST the limit is approached, which the limit alone does not.
b) CA(t)<2 means 40t<2(t2+16), since t2+16>0, that is t2−20t+16>0. The roots of t2−20t+16 are t=220±400−64=10±84=10±221, about 0.83 and 19.17. The quadratic is positive outside its roots, so CA(t)<2 for t<0.83 and for t>19.17. The first interval is the first minutes, while the concentration is still RISING past 2 mg/L; the question asks when it STAYS below, so the answer is from t=10+221≈19.17 h on. Keeping only one root, or the wrong one, is the usual loss here.
c) For t>0, t2+25=t1+t225, so CB(t)=1+t22520→20 mg/L. Never reached: t<t2+25 for every t, so t2+25t<1 and CB(t)<20. Next, CB(t)=16 means 20t=16t2+25; both sides are positive for t>0, so squaring is safe: 400t2=256(t2+25), 144t2=6400, t2=9400 and t=320≈6.67 h. Squaring is safe only because both sides are positive: always say so, since squaring can create solutions that the original equation does not have.
d) CB(t)>19.5 means 20t>19.5t2+25; squaring (both sides positive), 400t2>380.25t2+9506.25, so 19.75t2>9506.25, t2>481.33 and t>21.94 h. A check by the dominant term: the conjugate gives 20−CB(t)=t2+2520(t2+25−t)=t2+25(t2+25+t)500, which behaves like 2t2500=t2250; at t=21.94 that is about 0.52, close to 0.5 as expected.
e) The limit of a sum is the sum of the limits: C(t)→0+20=20 mg/L. For the ratio, CB(t)CA(t)=t2+1640t⋅20tt2+25=t2+162t2+25, which behaves like t22t=t2→0: in the long run the injection's share of the total becomes negligible. The side: C(t)−20=CA(t)−(20−CB(t)), which behaves like t40−t2250=t240t−250, positive for large t. So C(t) overshoots the steady state and comes back DOWN to 20 from above: at t=50, CA≈0.795 and CB≈19.901, total ≈20.70 mg/L. Comparing t1 with t21 is a comparison of powers, the only kind this chapter needs.
I tutor first-year calculus at Concordia and McGill, in English or in French, in Montreal or online, with the algebra review that goes with it. Get in touch for a first session.