Differential Calculus for Social Science 201-103 • CEGEP, Montreal
Corrected exercises: trigonometric functions and their derivatives (201-103)
These ten corrected exercises cover the trigonometric functions of differential calculus in the Social Science program: the circle and the identities as a refresher, the fundamental limit, the derivatives of sine, cosine and tangent, the chain rule on a periodic model and the study of a function over one period.
Exercise 0.1 opens the set with ten derivatives of sine, cosine and tangent to compute directly with the formulas, as in textbook lists: composites such as cos(2t+3π), tan(4t) and sin(t), products such as t2sint and sin(2t)cos(3t), the quotient 1+costsint, the power cos3t and a second derivative, each with a value to find at the given point and checked as you type it.
The thread of the set fits in one sentence: here sine is not a triangle tool, it is the MODEL of everything that comes back, and its derivative answers the question an economist or a demographer really asks, at what time of year is growth fastest?
Three pitfalls are named explicitly: differentiating in degrees, which adds a factor of pi over 180 to every formula (exercises 2 and 8); forgetting the factor left by the inner layer of a periodic model (exercise 6); and confusing the amplitude with the maximum value (exercises 4 and 8).
Self-checking setType your answers and check them question by 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, and your right answers stay filled in when you come back.
•Addition: sin(a+b)=sinacosb+cosasinb. Double angle: sin2x=2sinxcosx and cos2x=cos2x−sin2x.
•Fundamental limit: x→0limxsinx=1 and x→0limx1−cosx=0. Both are 00 forms.
•x→0limbxsinax=ba: it is the RATIO of the coefficients, never 1.
•Table: (sinx)′=cosx, (cosx)′=−sinx, (tanx)′=cos2x1=1+tan2x≥1, so the tangent function has no horizontal tangent line.
•Periodic model: f(t)=Asin(ωt+φ)+k. Amplitude A=2max−min, midline k=2max+min, period T=ω2π.
•The maximum is k+A, NOT A: the amplitude is measured from the midline (the axis of oscillation), not from zero.
•Derivative of a periodic model: f′(t)=Aωcos(ωt+φ). The factor ω comes from the inner layer and it is the number one mistake of the chapter.
•UNITS: ω is in radians per unit of time, so Aω is in units of f per unit of time. It is the quickest check.
•f′′(t)=−ω2(f(t)−k): the acceleration is proportional to the distance from the mean and has the opposite sign. This is the analytic definition of a cycle.
•Growth is MAXIMAL one quarter of a period before the peak, where the curve crosses its mean and where f′′ vanishes. The peak is the moment when f′=0.
•Two cycles of the same period add up to ONE single cycle: asinu+bcosu=Rsin(u+φ) with R=a2+b2 and tanφ=ab.
Calculations: the trigonometric derivatives to compute
Exercise 0.1: Differentiating sine, cosine and tangent: ten direct calculations
For each function, compute the derivative asked for using the formulas, then evaluate it at the point given. Angles are in radians.
One method for all ten lines: name the LAST operation (sum, product, quotient, power or composite function), apply its rule, then multiply by the derivative of the inside at every layer: (sinu)′=u′cosu, (cosu)′=−u′sinu, (tanu)′=cos2uu′. Values are read off the unit circle; the calculator, in radian mode, is only there to check.
a) f(t)=6sin(3t), then f′(3π).
b) g(t)=cos(2t+3π), then g′(12π).
c) h(t)=tan(4t), then h′(16π).
d) f(t)=2sint−3cost, then f′(6π).
e) g(t)=t2sint, then g′(2π).
f) h(t)=1+costsint, then h′(3π).
g) f(t)=cos3t, then f′(4π).
h) g(t)=sin(t), then g′(9π2).
i) h(t)=sin(2t)cos(3t), then h′(6π).
j) f(t)=5sin(2t): compute the SECOND derivative, then f′′(12π).
The move that decides everything is made BEFORE differentiating: name the last operation. cos3t is a power, sin(t) a composite function, t2sint a product, 1+costsint a quotient. Then every layer leaves its inside derivative as a factor, and that factor is the one people forget.
a)
f′(t)=6cos(3t)⋅3=18cos(3t). At t=3π, 3t=π and cosπ=−1: f′(3π)=−18. Pitfall: forgetting the 3 from the inside and answering −6.
b)
g′(t)=−sin(2t+3π)⋅2=−2sin(2t+3π). At t=12π, the angle is 6π+3π=2π and sin2π=1: g′(12π)=−2. The phase shift 3π is a constant: it leaves no factor.
c)
h′(t)=cos2(4t)4. At t=16π, 4t=4π and cos24π=21: h′(16π)=214=8. Check with the other form: 4(1+tan24π)=4×2=8.
d)
f′(t)=2cost−3⋅(−sint)=2cost+3sint. At t=6π: f′(6π)=2⋅23+3⋅21=3+1.5≈3.232. Pitfall: the sign, −3cost differentiates to +3sint.
e)
Product: g′(t)=2tsint+t2cost. At t=2π, sin=1 and cos=0: g′(2π)=2⋅2π=π≈3.142. Pitfall: writing (t2sint)′=2tcost, by multiplying together the derivatives of the two factors.
f)
Quotient: h′(t)=(1+cost)2cost(1+cost)−sint⋅(−sint)=(1+cost)2cost+cos2t+sin2t=(1+cost)21+cost=1+cost1, thanks to sin2t+cos2t=1. At t=3π: h′(3π)=1.51=32≈0.667.
g)
Power first, f(t)=(cost)3: f′(t)=3cos2t⋅(−sint)=−3cos2tsint. At t=4π: f′(4π)=−3⋅21⋅22=−432≈−1.061. Pitfall: confusing cos3t and cos(t3).
h)
g′(t)=cos(t)⋅2t1. At t=9π2, t=3π and cos3π=21: g′(9π2)=21⋅32π1=4π3≈0.239.
i)
Product of two composites: h′(t)=2cos(2t)cos(3t)+sin(2t)⋅(−3sin(3t))=2cos(2t)cos(3t)−3sin(2t)sin(3t). At t=6π, 2t=3π and 3t=2π: the first term vanishes because cos2π=0, and h′(6π)=−3⋅23⋅1=−233≈−2.598.
j)
f′(t)=10cos(2t), then f′′(t)=−20sin(2t)=−4f(t): the second derivative of a sine with angular frequency 2 is −22 times the function. At t=12π, sin6π=21: f′′(12π)=−10. Pitfall: a single factor 2 instead of two, which gives −5.
Overall check: for each line, the rate 0.002f(a+0.001)−f(a−0.001) computed by machine, in RADIAN mode, lands within one thousandth of the value found. In degree mode everything is off by an unrelated factor: it is the first setting to check.
Tick the exercises you have done or want to review: a free account, no password, keeps your ticks and your right answers from one visit to the next, tells you which chapter to tackle next and lets you ask for the exercise you are missing. Create your space, an email is enough.
Part A: the basics (/50)
Exercise 1: The unit circle, the radian and the identities: a refresher
This first exercise does not reopen the high school course. It gathers in ten minutes the only four things differential calculus needs: the unit circle, the special values, the Pythagorean identity and the related angles.
One idea is new, and it decides the rest of the chapter: here an angle is measured in RADIANS. A radian is a ratio of two lengths, hence a number without a unit, and that is what allows us to write sint where t is a number of months.
a) Convert 150∘ to radians, then 67π to degrees.
b) Give the exact values of cos32π, sin45π and tan3π.
c) A satisfaction index satisfies sinθ=53, with θ in the second quadrant. Find cosθ and tanθ.
d) Simplify sin(π−x), cos(π+x) and sin(−x).
e) A statistics package returns sin(30)=−0.988. Explain this result, then say what to type to get sin(30∘).
c)cosθ=−54=−0.8 and tanθ=−43=−0.75; the positive root is rejected.
d)sin(π−x)=sinx, cos(π+x)=−cosx, sin(−x)=−sinx.
e)The software computed the sine of 30 RADIANS. Type sin(30×180π)=0.5.
a)
The bridge between the two units is 180∘=π rad. So 150∘=150×180π=65π≈2.618 rad. In the other direction, 67π=67×180∘=210∘. Sanity check: 150∘ is past the right angle but short of the straight angle, so its measure in radians must fall between 2π≈1.571 and π≈3.142, which 2.618 satisfies. The classic pitfall is to multiply by π180 when you should divide: you can always catch it by checking that the result lands in the right interval. A botched conversion does not cost one mark, it spoils the whole model that follows.
b)
32π=120∘ is in the second quadrant, where cosine is NEGATIVE; its reference angle is 3π, so cos32π=−21=−0.5. 45π=225∘ is in the third quadrant, where sine is negative, with 4π as reference: sin45π=−22≈−0.7071. Finally tan3π=cos(π/3)sin(π/3)=1/23/2=3≈1.7321. The method comes down to two moves, and only one of them is a calculation: find the reference angle, then attach the sign of the quadrant.
c)
The Pythagorean identity sin2θ+cos2θ=1 gives cos2θ=1−259=2516, so cosθ=±54. The quadrant decides: in the second quadrant cosine is negative, so cosθ=−54=−0.8. Then tanθ=cosθsinθ=−4/53/5=−43=−0.75. The pitfall is costly: forgetting the sign amounts to placing the terminal point in the first quadrant, hence answering 0.8 and 0.75, two values that look right and are both wrong. The identity gives the SQUARE, never the sign.
d)
These three identities can be read off the figure without computing anything. π−x is the mirror image of x in the vertical axis: the y-coordinate stays, the x-coordinate changes sign, so sin(π−x)=sinx. π+x is the reflection through the centre: both coordinates change sign, so cos(π+x)=−cosx. Finally −x is the reflection in the horizontal axis, so sin(−x)=−sinx: sine is odd, cosine is even. These three equalities come back in exercise 5, where they are used to prove a derivative.
e)
The software read 30 RADIANS, not 30 degrees. Check: 30−4×2π≈30−25.133=4.867 rad, an angle in the fourth quadrant, where sine is negative; and indeed sin(4.867)≈−0.988. To get sin(30∘), type sin(30×180π)=sin6π=0.5. Remember the warning rule: every calculating package, every spreadsheet and every programming language works in radians by default. A report that announces a falling index where the phenomenon is rising almost always comes from this confusion of units.
Exercise 2: The fundamental limit: sine over x
At x=0, the quotient xsinx gives 00: an indeterminate form, just like those of the chapter on limits. Its value is neither 0 nor infinity, and it cannot be guessed.
It equals 1, and this is where ALL of the course's trigonometric differentiation comes from. The figure shows why: near zero, the sine curve and the line y=x become indistinguishable.
a) Compute 0.2sin(0.2) then 0.05sin(0.05), to four decimal places.
b) Give x→0limxsinx and say what the figure shows about this limit.
c) Show that x→0limx1−cosx=0 by multiplying by the conjugate, then check at x=0.01.
d) Compute x→0lim3xsin5x then x→0limsin3xsin5x.
e) Redo a) with a calculator set to DEGREES: what is 0.1sin(0.1∘), what limit does the quotient approach, and why does this rule out differentiating in degrees?
Show the solution
Answers
a)0.9933 then 0.9996.
b)x→0limxsinx=1: near zero sinx≈x, and the two curves of the figure merge.
c)The limit is 0; at x=0.01 the quotient is 0.0050.
d)Both limits equal 35≈1.6667.
e)0.01745, and the limit is 180π: in degrees (sinx)′=180πcosx, so we never differentiate in degrees.
a)
sin(0.2)=0.198669, so 0.2sin(0.2)=0.993347≈0.9933. Then sin(0.05)=0.0499792, so 0.05sin(0.05)=0.999583≈0.9996. The trend is clear: dividing the input by 4 divided the GAP FROM 1 by about 16. That is the signature of convergence in x2, and the approximation sinx≈x−6x3 confirms it, since it gives xsinx≈1−6x2. Careful: the calculator must be in radians, otherwise these two numbers are about 0.0175 and nothing converges to 1.
b)
x→0limxsinx=1. On the figure, the line y=x and the curve y=sinx visibly part ways from x=1 on, but stay glued together between −0.3 and 0.3: saying that the quotient tends to 1 means that sine and its argument become proportional with a ratio of 1, in other words that sinx≈x near zero. Geometrically, it is the arc merging with its chord as the angle gets small, and this is true ONLY if the angle is measured in radians, that is, by an arc length.
c)
Multiply top and bottom by the conjugate 1+cosx: x1−cosx=x(1+cosx)(1−cosx)(1+cosx)=x(1+cosx)1−cos2x=x(1+cosx)sin2x=xsinx×1+cosxsinx. The first factor tends to 1, the second to 20=0, so the product tends to 0. Check at x=0.01: 1−cos(0.01)=0.0000499996, divided by 0.01 this gives 0.0050. The conjugate method is the one from the chapter on algebraic limits; only the Pythagorean identity changes.
d)
For the first, force the same argument on top and bottom: 3xsin5x=35×5xsin5x, and since 5x→0, the second factor tends to 1: the limit is 35≈1.6667. For the second, build two reference quotients: sin3xsin5x=5xsin5x×sin3x3x×35, which gives 1×1×35≈1.6667. Remember the rule: when the two sines carry different coefficients, the limit is the RATIO of those coefficients, never 1. Answering 1 by reflex is the most common mistake on this page.
e)
0.1∘=0.1×180π=0.0017453 rad, whose sine is 0.0017453. The quotient is therefore 0.10.0017453=0.017453, and the limit is 180π≈0.01745, not 1. Direct consequence: in degrees, the derivative of sine is no longer cosine but 180πcosx, and this stray factor would contaminate every formula of the chapter, every chain rule and every optimization problem. The radian is not a teacher's convention: it is the only unit that keeps the table of derivatives clean.
Exercise 3: Sine, cosine and tangent as functions of time
In Social Science, sine is not a triangle tool: it is THE model of everything that comes back. Retail sales, emergency room traffic, seasonal migration, the circadian rhythm: all are described by f(t)=Asin(ωt+φ)+k.
Four letters, four management questions. A is the amplitude, half the gap between the best and the worst moment. ω sets the period, T=ω2π. φ shifts the cycle along the calendar. k is the average level around which everything oscillates.
a) For f(t)=8sin(6πt)+30, give the amplitude, the period, the midline (axis of oscillation), the maximum and the minimum.
b) For g(t)=12sin(6πt+3π)+40, rewrite in the form Asin(ω(t−t0))+k, give t0, then the first t≥0 at which g reaches its maximum.
c) Give the period of the tangent and say why it cannot model any bounded seasonal phenomenon.
d) The three other trigonometric functions are secant, cosecant and cotangent: say which of the three basic functions each one is the reciprocal of.
e) Compare the RELATIVE amplitude kA of f and of g, as a percentage, and say which of the two series is more seasonal.
Show the solution
Answers
a)A=8, T=12, midline y=30, maximum 38, minimum 22.
b)g(t)=12sin(6π(t+2))+40, so t0=−2; first maximum at t=1, where g=52.
c)Period π. The tangent is unbounded and has vertical asymptotes, so it models no bounded cycle.
e)26.7 percent against 30 percent: g is the more seasonal.
a)
Read directly: A=8, so the amplitude is 8 units. ω=6π, so the period is T=π/62π=12 time units, which makes it the standard annual model when t counts months. The midline is the line y=k=30. The maximum is k+A=38, the minimum k−A=22. Check: the average of the two is indeed 238+22=30, and their half-difference is 238−22=8. These two formulas, k=2max+min and A=2max−min, will be used in reverse in exercise 4, where the data supply the maximum and the minimum.
b)
Factor ω out of the argument: 6πt+3π=6π(t+2), so g(t)=12sin(6π(t+2))+40 and t0=−2. The shift is NEGATIVE, which means the cycle runs two months ahead of the model without a phase shift. The maximum is reached when the argument equals 2π: 6π(t+2)=2π gives t+2=3, so t=1, and g(1)=52. The pitfall is to read t0=+2 because the statement shows a plus: the standard form is written t−t0, so a plus in the argument means a negative t0, exactly as for the parabola.
c)
The tangent has period π, not 2π: it repeats its values twice as fast as sine, because the simultaneous sign change of sine and cosine leaves their quotient unchanged. Above all, it has no amplitude: it starts at −∞ and climbs to +∞ on each interval, with a vertical asymptote at 2π+kπ. A real seasonal phenomenon is bounded, sales do not become infinite in December, so the tangent is never used as a cyclical model. It appears in this course for another reason, as a result of differentiation and in the equations of the type tanu=c of exercise 10.
d)
Secant is the reciprocal of cosine, secx=cosx1; cosecant is the reciprocal of sine, cscx=sinx1; cotangent is the reciprocal of tangent, cotx=sinxcosx. Watch the vocabulary: the prefix co does not carry over, CO-secant answers to SINE. These three functions appear only in some course outlines and will not be used further here, where everything is written with the three basic functions.
e)
kA measures the weight of the season relative to the usual level, and it is this proportion, not the raw amplitude, that a manager compares from one series to another. For f: 308=0.2667, or 26.7 percent. For g: 4012=0.30, or 30 percent. Series g is therefore the more seasonal, even though both have the same period. Comparing the amplitudes alone, 8 against 12, would give the same conclusion here by chance; it would flip as soon as the average levels diverge, for instance between an amplitude of 12 on a level of 200 and an amplitude of 8 on a level of 30.
Exercise 4: Reading the four parameters off twelve months of data
A retail chain records its monthly sales, in thousands of dollars, over a full year. The statistics department has fitted a sinusoid f(t)=Asin(ωt+φ)+k to the scatter plot, where t is the month number, from 1 for January to 12 for December.
The fitted maximum is 140 in December and the minimum is 80 in June. The whole exercise consists in working back from the data to the four parameters, then translating them into sentences for a report.
a) Deduce A and k from the maximum and the minimum.
b) Give the period of the model in months, then ω in radians per month, to four decimal places.
c) Find φ knowing that the maximum is reached at t=12, then write f(t).
d) Predict the sales for September, t=9, then compare with the observed value of 112.
e) Translate k, A and 2A into one report sentence each.
Show the solution
Answers
a)A=30 and k=110 thousand dollars.
b)T=12 months and ω=6π≈0.5236 rad/month.
c)φ=2π≈1.5708, so f(t)=30sin(6πt+2π)+110.
d)f(9)=110 thousand dollars, against 112 observed: a gap of 2 thousand, or 1.8 percent.
e)k: 110 thousand a month on average. A: a seasonal gap of 30 thousand from the average. 2A: 60 thousand between December and June.
a)
The chapter's two formulas are used here in reverse from exercise 3: A=2max−min=2140−80=30 and k=2max+min=2140+80=110, both in thousands of dollars. The pitfall is to take A=140, that is, to confuse the amplitude with the maximum value: the amplitude is measured from the AXIS, not from zero. This confusion returns in exercise 8, statement d), and it then throws the derivative off by a factor of 4 in this example.
b)
The phenomenon is annual and the data are monthly, so the period is T=12 months. From this, ω=T2π=122π=6π≈0.5236 radians per month. The units deserve a second look: ω is not a pure number, it is a number of radians PER MONTH, and that is what will make Aω a number of thousands of dollars per month in exercise 6. A student who writes ω=12 has swapped the period and the angular frequency; the curve would then oscillate about twice a month.
c)
The maximum of a sine is reached when its argument equals 2π modulo 2π. At t=12: 6π×12+φ=2π+φ, which we want equal to 2π+2π, so φ=2π≈1.5708. The model is written f(t)=30sin(6πt+2π)+110, which is also f(t)=110+30cos(6πt), since a sine shifted by a quarter turn IS a cosine. Both forms are correct; keeping the first avoids switching functions in the middle of the chapter.
d)
f(9)=30sin(23π+2π)+110=30sin(2π)+110=110 thousand dollars. The observed value is 112, a gap of 2 thousand, about 1.8 percent of the observed value. That is the right order of magnitude for a seasonal fit: the model is a SMOOTHING, not a prophecy, and a gap of a few percent in a given month invalidates nothing. What must be checked is that the gaps change sign from one month to the next instead of piling up on the same side, a sign of an underlying trend that the sinusoid alone does not capture.
e)
k=110: on average over the year, the chain sells 110 thousand dollars a month. A=30: the best month exceeds this average by 30 thousand, and the worst falls short of it by the same amount. 2A=60: between December and June, the sales gap reaches 60 thousand dollars a month, more than half the average level. This third number drives the cash-flow plan and seasonal hiring, and it is the one people forget to compute when they stop at the amplitude.
Exercise 5: The derivatives of sine, cosine and tangent
Nobody memorizes three derivatives: we establish ONE, the derivative of sine, and the other two follow in a few lines. The proof needs only two things already in hand, the angle addition identity of exercise 1 and the fundamental limit of exercise 2.
The differentiation rules themselves are those of the previous chapter and do not change: product, quotient, chain. Only the table of basic functions grows by three lines.
a) Starting from sin(x+h)=sinxcosh+cosxsinh, show that hsin(x+h)−sinx=cosx⋅hsinh−sinx⋅h1−cosh, then conclude about (sinx)′.
b) Deduce (cosx)′ by writing cosx=sin(2π−x).
c) Establish (tanx)′=cos2x1=1+tan2x by the quotient rule, then say why the tangent function never has a horizontal tangent line.
d) Compute f′(3π) for f=sin, g′(6π) for g=cos and h′(4π) for h=tan.
e) Let p(x)=sinxcosx. Differentiate by the product rule, then by noting that p(x)=21sin2x. What does the comparison of the two answers confirm?
Show the solution
Answers
a)(sinx)′=cosx: the first quotient tends to 1, the second to 0.
b)(cosx)′=−sinx, the sign coming from the derivative of 2π−x.
c)(tanx)′=cos2x1=1+tan2x≥1, so never zero: no horizontal tangent line.
d)0.5, then −0.5, then 2.
e)cos2x−sin2x on one side, cos2x on the other: the comparison gives back the double-angle identity.
a)
Replace sin(x+h) by the angle addition identity: hsinxcosh+cosxsinh−sinx=hcosxsinh−hsinx(1−cosh)=cosx⋅hsinh−sinx⋅h1−cosh. As h→0, the first quotient tends to 1 and the second to 0, by the two limits of exercise 2. What remains is cosx×1−sinx×0=cosx, so (sinx)′=cosx. Note that x stayed a SPECTATOR throughout the passage to the limit: it is h that tends to zero, and that is why cosx and sinx come out of the quotients like constants.
b)
cosx=sin(2π−x) is the cofunction identity. The chain rule gives (cosx)′=cos(2π−x)×(−1)=−sinx, since cos(2π−x)=sinx. The minus sign comes from the derivative of the inner layer, 2π−x, and not from a convention to memorize. Here is the exact statement of the table: the derivative of SINE is cosine, with no sign; the derivative of COSINE is minus sine. Swapping the two is statement a) of exercise 8.
c)
tanx=cosxsinx, so the quotient rule gives (tanx)′=cos2xcosxcosx−sinx(−sinx)=cos2xcos2x+sin2x=cos2x1 by the Pythagorean identity. Dividing term by term instead gives the second form, 1+tan2x. Both are useful: the first for evaluating, the second for solving. Immediate consequence: cos2x1≥1 always, so the derivative of the tangent NEVER vanishes and the curve has no horizontal tangent line. This is the analytic translation of the fact that the tangent function is strictly increasing on each of its intervals.
d)
f′(x)=cosx, so f′(3π)=cos3π=21=0.5. g′(x)=−sinx, so g′(6π)=−sin6π=−21=−0.5. h′(x)=cos2x1, so h′(4π)=(2/2)21=1/21=2. Graphical reading: at 3π the sine curve is still rising, but half as fast as at the origin; at 6π cosine is already falling. The sign can be checked by eye on any graph, and it is the quickest check.
e)
By the product rule: p′(x)=cosxcosx+sinx(−sinx)=cos2x−sin2x. By the second form: p(x)=21sin2x, and the chain rule gives p′(x)=21×2cos2x=cos2x. The two answers are necessarily equal, so cos2x=cos2x−sin2x: the differentiation has just RECOVERED the double-angle identity. Check at x=6π: cos3π=0.5 on one side, 43−41=0.5 on the other. Remember the move: when two paths lead to the same derivative, their equality is a free identity, and it is often the simplest of proofs.
Part B: problems and reasoning (/50)
Exercise 6: The chain rule on a periodic model
We return to the sales model fitted in exercise 4: f(t)=30sin(6πt+2π)+110, in thousands of dollars, t in months.
Differentiating such a model means applying the chain rule to an affine inner layer, and the chapter's number one mistake fits in one word: forgetting the factor ω left by that layer.
a) Give f′(t) for the general model f(t)=Asin(ωt+φ)+k, and say what happens to each of the four letters.
b) Apply this to the sales model: give f′(t) then the amplitude of f′, to two decimal places.
c) Compute f′(3) and f′(9), to two decimal places, and interpret each in one report sentence.
d) Give the units of f′, then give as a percentage the error of a student who forgets the factor ω.
e) Show that f′′(t)=−ω2(f(t)−k) and say what this equality tells us about a cyclical phenomenon.
Show the solution
Answers
a)f′(t)=Aωcos(ωt+φ): A stays, ω is added as a factor, φ does not move, k disappears.
b)f′(t)=5πcos(6πt+2π), with amplitude 5π≈15.71 thousand per month.
c)f′(3)≈−15.71: in March sales fall by 15.71 thousand per month. f′(9)≈+15.71: in September they rise by the same amount.
d)Thousands of dollars per month; forgetting ω overestimates the rate by 91 percent, a factor of π6.
e)f′′(t)=−ω2(f(t)−k)≈−0.2742(f(t)−110): the acceleration pulls the series back toward its mean.
a)
The outer layer is sine, the inner layer is u(t)=ωt+φ, whose derivative is ω. The chain rule therefore gives f′(t)=Aωcos(ωt+φ). Each letter has its own fate: A stays as a factor, ω appears IN ADDITION as a factor, φ stays in its place inside the argument, and k disappears, since the derivative of a constant is zero. This last point deserves to be said out loud: the average level has no effect on the rate of change, and two series with very different levels can have exactly the same derivative.
b)
Here A=30, ω=6π and φ=2π, so f′(t)=30×6πcos(6πt+2π)=5πcos(6πt+2π). The amplitude of f′ is Aω=5π≈15.71. In other words, the rate of change of sales oscillates between −15.71 and +15.71 thousand dollars per month, never beyond. This number is the most useful in the whole model for a manager: it is the maximum monthly change to prepare for, in either direction.
c)
f′(3)=5πcos(2π+2π)=5πcosπ=−5π≈−15.71. Report sentence: in March, sales are falling at a rate of 15.71 thousand dollars per month. f′(9)=5πcos(23π+2π)=5πcos2π=5π≈+15.71: in September, they are rising at a rate of 15.71 thousand dollars per month. These two months are exactly those where the model crosses its mean, f(3)=f(9)=110: the level is ordinary there, the rate of change is extreme. This is the thread of the whole set, and exercise 7 makes it its central question.
d)
f is measured in thousands of dollars and t in months, so f′ is measured in THOUSANDS OF DOLLARS PER MONTH, the unit of the output divided by that of the input. A student who forgets ω writes 30cos(⋯) and announces an amplitude of 30 instead of 5π≈15.708: the rate is overestimated by a factor of 5π30=π6≈1.91, an error of 91 percent. On a cash-flow plan, that amounts to setting aside nearly double the real need. Units do not save us here, because both answers are dimensionally consistent: only the chain rule protects.
e)
Differentiate a second time: f′′(t)=−Aω2sin(ωt+φ). But f(t)−k=Asin(ωt+φ), so f′′(t)=−ω2(f(t)−k). For the sales model, ω2=36π2≈0.2742, hence f′′(t)≈−0.2742(f(t)−110). The equality says exactly what cyclical means: the acceleration is proportional to the DISTANCE FROM THE MEAN, with the opposite sign. The further the series strays from its usual level, the stronger the pull that brings it back; and when it crosses its mean, the acceleration is zero, so the rate of change is at its maximum there. It is the same equation that governs every return to equilibrium.
Exercise 7: The peak is not the month when sales rise fastest
Here is the question a manager really asks, and the one only the derivative answers: at what time of year are sales increasing fastest? It is not at the peak, where they have finished rising.
We keep working with f(t)=30sin(6πt+2π)+110, on the interval [0,12].
a) Solve f′(t)=0 on [0,12] and say what kind of point each solution is.
b) Compute f′′(t), then solve f′′(t)=0 on [0,12] and give the value of f at these points.
c) Give f′(3) and f′(9) and say which of the two months is the one when sales rise fastest.
d) State the general rule: how far from the peak is the moment of maximum growth?
e) Draw the sign chart of f on [0,12] and write the management sentence that follows from it.
Show the solution
Answers
a)t=0, t=6 and t=12; t=6 is a minimum, f(6)=80, and the two endpoints are maxima, f=140.
b)f′′(t)=−65π2cos(6πt), zero at t=3 and t=9, where f=110.
c)f′(3)≈−15.71 and f′(9)≈+15.71: September is when sales rise fastest.
d)One quarter of a period before the peak, which is 3 months here.
e)f decreases on [0,6] from 140 to 80, then increases to 140; stock is ordered on December, hiring on September.
a)
f′(t)=5πcos(6πt+2π), which is also −5πsin(6πt). It vanishes when sin(6πt)=0, so when 6πt=nπ, that is t=6n: on [0,12], the solutions are t=0, t=6 and t=12. The sign of f′ decides their nature: f′ is negative on (0,6) and positive on (6,12), so t=6 is a MINIMUM, with f(6)=80, while t=0 and t=12 are maxima, with f=140. These are June and December, as exercise 4 already announced.
b)
f′′(t)=−30(6π)2sin(6πt+2π)=−65π2cos(6πt)≈−8.225cos(6πt). It vanishes when cos(6πt)=0, so when 6πt=2π+nπ, that is t=3+6n: on [0,12], t=3 and t=9. The second derivative changes sign at each, so these are two inflection points, and f(3)=f(9)=110: they fall exactly on the midline. Remembering this saves a calculation every time: a sinusoid always bends where it crosses its mean.
c)
From exercise 6, f′(3)=−5π≈−15.71 and f′(9)=+5π≈+15.71. The month when sales rise fastest is therefore SEPTEMBER, t=9, and the month when they fall fastest is March. Yet in September the level of sales is perfectly ordinary, 110 thousand dollars, exactly the annual average: a dashboard that looks only at the level sees nothing happening that month. This is precisely what the derivative brings, and what no reading of the sales column gives.
d)
The maximum of f′ is reached ONE QUARTER OF A PERIOD before the peak of f. Here the period is 12 months, so a quarter is 3 months, and the peak is in December: maximum growth falls in September, 12−3=9. The rule comes from the very structure of the model: f′=Aωcos(ωt+φ) is the same sinusoid moved ahead by a quarter turn, since cosu=sin(u+2π). It holds for any cyclical phenomenon, whatever the period, and it is what we will find again over seven days in exercise 9, then on a two-cycle model in exercise 10.
e)
Sign chart on [0,12]: tf′f00140−↘6080+↗120140 with two inflection points, at t=3 and t=9, where f equals 110. The management sentence: stock and floor staffing are ordered on the LEVEL, hence on December; seasonal hiring and ramping up the warehouse are triggered on the RATE OF CHANGE, hence as early as September. Confusing the two means hiring at the moment when growth is already over.
Exercise 8: Five statements to correct
Each of the following five statements is FALSE. Say why, then give the correct statement.
All five come from real student papers, and each one costs the whole question where it appears, because it spoils everything that follows.
a) “The derivative of sine is minus cosine.”
b) “The period of sin(3t) is 6π.”
c) “Since the calculator displays degrees, we differentiate in degrees.”
d) “The amplitude of a model Asin(ωt+φ)+k is its maximum value.”
e) “xsinx tends to 1 as x tends to infinity.”
Show the solution
Answers
a)False: (sinx)′=cosx; the minus sign belongs to (cosx)′=−sinx.
b)False: the period is 32π≈2.094; 6π would be that of sin(3t).
c)False: we differentiate in radians; in degrees (sinx)′=180πcosx.
d)False: the amplitude is A and the maximum k+A, which is 140 for A=30 and k=110.
e)False: the limit is 1 at 0 and 0 at +∞, by the squeeze theorem.
a)
False: (sinx)′=cosx, with no sign at all. The minus sign belongs to the other line of the table, (cosx)′=−sinx. The confusion can be spotted by eye on a graph: at x=0 the sine curve RISES, so its derivative is positive there, and cos0=1 fits whereas −cos0=−1 is absurd. Correct statement: the derivative of sine is cosine, the derivative of cosine is minus sine. What costs dearly is not the memory slip but its knock-on effect: on a seasonal model, it swaps the month of maximum growth and the month of maximum decline, so it reverses the decision.
b)
False: the period of sin(ωt) is ω2π, so that of sin(3t) is 32π≈2.094. The coefficient DIVIDES the period, it does not multiply it: multiplying the argument by 3 makes the phenomenon cycle three times faster. The value 6π would be that of sin(3t). Direct check: sin(3(t+32π))=sin(3t+2π)=sin(3t), the expected equality. Correct statement: the period of sin(3t) is 32π.
c)
False, and it is the most costly of the five. The whole table of derivatives rests on x→0limxsinx=1, a limit that is true only in radians; in degrees it equals 180π≈0.01745, as exercise 2 measured. In degrees, the table would become (sinx)′=180πcosx, and this stray factor would spread into every chain rule and every optimization. Correct statement: in differential calculus we work in radians, and the calculator display is set accordingly before the first calculation.
d)
False: the amplitude is A, the HALF-difference between the maximum and the minimum, and it is measured from the midline y=k, not from zero. The maximum value is k+A and the minimum value k−A. On the sales model, A=30 and k=110 give a maximum of 140, not 30. Correct statement: A=2max−min and max=k+A. The error is doubly costly, because it also throws off the derivative: the amplitude of the rate of change is Aω, so taking 140 for A would multiply the announced rate by more than four.
e)
False: the limit 1 is the one at ZERO, not at infinity, and the two are unrelated. At infinity, the numerator stays bounded between −1 and 1 while the denominator grows without bound, so xsinx≤x1→0: by the squeeze theorem, x→+∞limxsinx=0. Check: at x=100, the quotient is about −0.0051. Correct statement: the limit is 1 at 0 and 0 at +∞. Keep the matching reading in mind: an oscillation divided by a growing time damps out, which is exactly the model of a cycle whose relative effect fades away.
Exercise 9: Problem: emergency room traffic over seven days
A hospital's records department models the number of visits recorded in the emergency room by N(t)=175+45cos(72πt), where t is the number of days elapsed since Monday at 12 a.m.
The cycle here is weekly rather than yearly, but the manager's questions do not change: which is the busiest day, and when does the load climb fastest?
a) Give the amplitude, the period, the mean, the maximum and the minimum of N.
b) Give N′(t), compute N′(2) to one decimal place and interpret it in one report sentence.
c) Solve N′(t)=0 on [0,7] and name the corresponding days.
d) Solve N′′(t)=0 on [0,7], then give the value of N′ at each of these times, to two decimal places.
e) Draw the sign chart on [0,7] and say which indicator, the level or the rate of change, each of the department's two decisions is based on.
Show the solution
Answers
a)A=45, T=7 days, mean 175, maximum 220 on Monday, minimum 130 on Thursday at noon.
b)N′(t)=−790πsin(72πt) and N′(2)≈−39.4: on Wednesday morning, traffic is falling by 39 visits per day.
c)t=0, t=3.5 and t=7: Monday for the maxima, Thursday noon for the minimum.
d)t=1.75 and t=5.25, where N′ equals −40.39 then +40.39 visits per day.
e)N decreases from 220 to 130, then climbs back to 220. The schedule is based on the level, hence Monday; the staff call-back on the rate of change, hence Saturday morning.
a)
The reading is direct: A=45 visits, k=175 visits, and ω=72π gives a period T=2π/72π=7 days, which confirms that the cycle is weekly. The maximum is k+A=220 visits and the minimum k−A=130 visits. The maximum is reached when cos(72πt)=1, so at t=0 and t=7, that is, on Monday; the minimum when cosine equals −1, so at t=3.5, that is, Thursday at noon. The gap between the worst and the best day reaches 2A=90 visits, almost half the mean: this is not a scheduling detail, it is a safety question.
b)
The chain rule gives N′(t)=−45×72πsin(72πt)=−790πsin(72πt)≈−40.39sin(72πt). At t=2, Wednesday at 12 a.m.: 74π≈1.7952 rad, whose sine is 0.9749, so N′(2)≈−40.39×0.9749≈−39.4. Report sentence: on Wednesday morning, daily traffic is falling at a rate of about 39 visits per day. The negative sign is the one expected, since the curve is going down from the Monday peak to the Thursday trough.
c)
N′(t)=0 when sin(72πt)=0, so when 72πt=nπ, that is t=27n. On [0,7]: t=0, t=3.5 and t=7. The two endpoints correspond to Monday and give the maximum, 220 visits; t=3.5 is Thursday noon and gives the minimum, 130 visits. We do not conclude from the derivative vanishing alone: it is the change of SIGN of N′, negative before 3.5 and positive after, that makes Thursday noon a minimum and not just a critical point.
d)
N′′(t)=−790π×72πcos(72πt)=−49180π2cos(72πt)≈−36.26cos(72πt). It vanishes when 72πt=2π+nπ, so t=1.75 and t=5.25 on the interval. At t=1.75, Tuesday late in the day, N′=−790π≈−40.39: this is the fastest drop of the week. At t=5.25, Saturday early in the morning, N′=+790π≈+40.39: this is the fastest recovery. As the quarter-period rule predicts, 5.25 comes before the Monday peak, t=7, by 1.75 days, exactly 47.
e)
Sign chart: tN′N00220−↘3.50130+↗70220 with two inflection points, at t=1.75 and t=5.25, where N=175. The two decisions do not read the same row. The baseline SCHEDULE, how many doctors to put on the roster, is based on the LEVEL: reinforcements are needed on Monday, when 220 visits are expected. The staff call-back PROTOCOL, on the other hand, is based on the RATE OF CHANGE: it must be triggered on Saturday morning, when the load is climbing by 40 visits per day, while the level is still perfectly average there, 175 visits. Waiting for the level to rise means arriving a day and a half too late.
Exercise 10: Problem: two superimposed cycles and the month of the peak
A retailer breaks down the gap between its sales and their annual mean into two seasonal effects of the same period: the school calendar and the weather. The research department obtains E(t)=5sin(6πt)+12cos(6πt), in thousands of dollars, with t the number of months elapsed since January 1.
Two cycles of the same period always add up to ONE SINGLE cycle of that period, shifted: it is the angle addition identity read backwards, and it is what makes the problem solvable by hand.
a) Write E(t)=Rsin(6πt+φ): give R exactly and φ to three decimal places.
b) Give the amplitude and the period of E, then the gap between the best and the worst month.
c) Differentiate E and solve E′(t)=0 on [0,12): give the two times to two decimal places.
d) Say which one is a maximum and which a minimum, with the value of E at each.
e) Find the time of year when the gap GROWS fastest, give this rate to two decimal places, and check the quarter-period rule.
Show the solution
Answers
a)R=13 and φ=arctan512≈1.176 rad.
b)Amplitude 13 thousand dollars, period 12 months, gap 2R=26 thousand.
c)E′(t)=6π(5cosu−12sinu), zero at t≈0.75 and t≈6.75 months.
d)t≈0.75 is a maximum, E=13; t≈6.75 is a minimum, E=−13.
e)Maximum growth at t≈9.75 months, at a rate of 613π≈6.81 thousand per month, which is 3 months before the peak at t≈12.75.
a)
Expand the target form: Rsin(u+φ)=Rcosφsinu+Rsinφcosu, with u=6πt. Matching gives Rcosφ=5 and Rsinφ=12. Squaring and adding, the Pythagorean identity makes φ disappear: R2=25+144=169, so R=13 exactly. Then tanφ=512 with a cosine and a sine that are both positive, so φ is in the first quadrant: φ=arctan512≈1.176 rad. The 5, 12, 13 triangle is no accident of the statement, it is what makes R an integer. Check at t=0: E(0)=12 on one side, 13sin(1.176)=12 on the other.
b)
The amplitude is R=13 thousand dollars, and the period stays π/62π=12 months: adding two cycles of the same period never changes the period, only the amplitude and the shift. The gap between the best and the worst month is 2R=26 thousand dollars. Notice that 13 is less than 5+12=17: the two effects do not fully reinforce each other, because they are out of phase by a quarter turn. Adding the amplitudes is the classic mistake here, and it overestimates the seasonality by 31 percent.
c)
With the expanded form, E′(t)=5×6πcos(6πt)−12×6πsin(6πt)=6π(5cosu−12sinu). It vanishes when 5cosu=12sinu, so when tanu=125, which gives u=arctan125≈0.3948 then u≈0.3948+π≈3.5364. Since t=π6u, we get t≈0.75 and t≈6.75 months, that is, around January 23 and around July 23. The factor 6π plays no role in the solving, but it will play one in the value of the rate in e): we do not throw it away, we set it aside.
d)
With the compact form, E(t)=13sin(u+φ). At t≈0.75: u+φ≈0.3948+1.176=1.5708=2π, so E=13, a MAXIMUM. At t≈6.75: u+φ≈3.5364+1.176=4.7124=23π, so E=−13, a MINIMUM. The compact form did all the work: looking at the sign of E′′ would have given the same answer in twice as many lines. Translation: sales exceed their annual mean by 13 thousand dollars around January 23, and fall short of it by the same amount around July 23.
e)
E′(t)=6π(5cosu−12sinu) can also be written E′(t)=613πcos(u+φ), by the same matching. This rate of change is at its maximum when cos(u+φ)=1, so when u+φ=2π, that is u≈6.2832−1.176=5.1072 and t=π6u≈9.75 months, around October 23. The rate there is 613π≈6.81 thousand dollars per month. Check of the quarter-period rule: the next peak is at t≈0.75+12=12.75, and 12.75−9.75=3, exactly one quarter of the 12 months. The rule therefore also holds on a composite model, as expected, since the sum of two cycles of the same period IS a sinusoid.
Something missing from this set?A type of exercise your teacher gives, a problem you are stuck on, a mistake you spotted: tell me, I read every request.
Sending goes through your free space: that is how I can answer you and tell you when the exercise is added. An email is enough, no password. Create your space.
Rather not create an account? Write to me directly.
Looking for a differential calculus tutor in Montreal?
Get in touch for a first session. Trigonometry is the chapter where Social Science students fall behind fastest, because nobody told them what it is for in their program.