Type a command or search the notes. Tab cycles completions. Up and down recall history when blank. Escape dismisses, then leaves the prompt.
MATH
[1251]

Seen questions

Worked solutions on trigonometric integrals, average values, reduction formulae, substitutions and partial fractions.

These past exam questions combine product-to-sum identities, tangent-secant integrals, reduction formulae, substitutions and partial fractions. The first chooses a method for each branch of a piecewise integral; the second combines a reduction formula with a substitution and partial fractions. In each problem, identify the structure that makes the method apply before starting the algebra and track interval and branch restrictions when changing variables.

[METHOD]

Product-to-sum

Convert the product to a sum, then integrate term by term:

sin⁡Acos⁡B=12(sin⁡(A+B)+sin⁡(A−B))\sin A\cos B=\tfrac12\bigl(\sin(A+B)+\sin(A-B)\bigr) cos⁡Acos⁡B=12(cos⁡(A+B)+cos⁡(A−B))\cos A\cos B=\tfrac12\bigl(\cos(A+B)+\cos(A-B)\bigr) sin⁡Asin⁡B=−12(cos⁡(A+B)−cos⁡(A−B))\sin A\sin B=-\tfrac12\bigl(\cos(A+B)-\cos(A-B)\bigr)
[METHOD]

Powers of tangent and secant

To evaluate ∫tan⁡mxsec⁡nx dx\int \tan^m x \sec^n x \, dx:

  • If nn is a positive even integer, factor out sec⁡2x\sec^2 x and substitute u=tan⁡xu=\tan x. Convert remaining powers of sec⁡\sec by 1+tan⁡2x=sec⁡2x1+\tan^2 x=\sec^2 x.
  • If n=0n=0, use tan⁡2x=sec⁡2x−1\tan^2 x=\sec^2 x-1 to lower the power of tan⁡\tan by two, with ∫dx\int dx and ∫tan⁡x dx\int\tan x\,dx as base cases.
  • If nn is odd:
    • If mm is odd, factor out tan⁡xsec⁡x\tan x\sec x and substitute u=sec⁡xu=\sec x. Convert remaining powers of tan⁡\tan by tan⁡2x=sec⁡2x−1\tan^2 x=\sec^2 x-1.
    • If mm is even, write tan⁡mx=(sec⁡2x−1)m/2\tan^m x=(\sec^2 x-1)^{m/2} and expand into powers of sec⁡x\sec x. Reduce odd secant powers by parts until only ∫sec⁡x dx\int\sec x\,dx remains.
[EXAMPLE]

Average value of a piecewise function

The average value of a function ff on [a,b][a,b] is

fˉ=1b−a∫abf(x) dx.\bar f=\frac{1}{b-a}\int_a^b f(x)\,dx.

Compute the average fˉ\bar f on [−π4,π4]\left[-\frac{\pi}{4},\frac{\pi}{4}\right] for

f(x)={cos⁡3xsin⁡2x,−π4≤x≤0,tan⁡3xsec⁡2x,0<x≤π4.f(x)= \begin{cases} \cos 3x\sin 2x,&-\frac{\pi}{4}\le x\le0,\\[2mm] \tan^3x\sec^2x,&0<x\le\frac{\pi}{4}. \end{cases}

Solution

The interval is split at x=0x=0, so the integral of ff is the sum of one integral from each branch. Parts (a) and (b) find the two antiderivatives, part (c) applies them on the appropriate intervals and part (d) asks whether ff attains its average.

(a) The product-to-sum integral

The product-to-sum identity

sin⁡Acos⁡B=12(sin⁡(A+B)+sin⁡(A−B))\sin A\cos B=\frac12\bigl(\sin(A+B)+\sin(A-B)\bigr)

applies after writing cos⁡3xsin⁡2x\cos3x\sin2x as sin⁡2xcos⁡3x\sin2x\cos3x. Thus

I=∫cos⁡3xsin⁡2x dx=12∫(sin⁡(2x+3x)+sin⁡(2x−3x)) dx=12∫(sin⁡5x−sin⁡x) dx=−cos⁡5x10+cos⁡x2+C.\begin{aligned} I &=\int \cos3x\sin2x\,dx\\ &=\frac12\int\bigl(\sin(2x+3x)+\sin(2x-3x)\bigr)\,dx\\ &=\frac12\int\bigl(\sin5x-\sin x\bigr)\,dx\\ &=-\frac{\cos5x}{10}+\frac{\cos x}{2}+C. \end{aligned}

The difference of arguments is −x-x and sin⁡(−x)=−sin⁡x\sin(-x)=-\sin x; this sign is easy to lose.

(b) The tangent-secant integral

Use the substitution

u=tan⁡x,du=sec⁡2x dx.u=\tan x, \qquad du=\sec^2x\,dx.

Then

J=∫tan⁡3xsec⁡2x dx=∫u3 du=u44+C=tan⁡4x4+C.J=\int\tan^3x\sec^2x\,dx =\int u^3\,du =\frac{u^4}{4}+C =\frac{\tan^4x}{4}+C.

For the definite integral on the second branch, the bounds transform as

x=0⟹u=tan⁡0=0,x=π4⟹u=tan⁡π4=1.x=0\Longrightarrow u=\tan0=0, \qquad x=\frac{\pi}{4}\Longrightarrow u=\tan\frac{\pi}{4}=1.

Therefore

∫0π/4tan⁡3xsec⁡2x dx=∫01u3 du=[u44]01=14.\int_0^{\pi/4}\tan^3x\sec^2x\,dx =\int_0^1u^3\,du =\left[\frac{u^4}{4}\right]_0^1 =\frac14.

(c) Computing the average

For the first branch, use the antiderivative from (a):

∫−π/40cos⁡3xsin⁡2x dx=[−cos⁡5x10+cos⁡x2]−π/40=25−3210.\begin{aligned} \int_{-\pi/4}^{0}\cos3x\sin2x\,dx &=\left[-\frac{\cos5x}{10}+\frac{\cos x}{2}\right]_{-\pi/4}^{0}\\ &=\frac25-\frac{3\sqrt2}{10}. \end{aligned}

The second branch contributes 14\frac14 by (b). Each formula for ff applies only on its own subinterval, so these two integrals give the total:

∫−π/4π/4f(x) dx=∫−π/40cos⁡3xsin⁡2x dx+∫0π/4tan⁡3xsec⁡2x dx=(25−3210)+14=13−6220.\begin{aligned} \int_{-\pi/4}^{\pi/4}f(x)\,dx &=\int_{-\pi/4}^{0}\cos3x\sin2x\,dx +\int_{0}^{\pi/4}\tan^3x\sec^2x\,dx\\ &=\left(\frac25-\frac{3\sqrt2}{10}\right)+\frac14\\ &=\frac{13-6\sqrt2}{20}. \end{aligned}

The interval length is

π4−(−π4)=π2.\frac{\pi}{4}-\left(-\frac{\pi}{4}\right)=\frac{\pi}{2}.

Consequently,

fˉ=1π/2⋅13−6220=13−6210π.{ \bar f =\frac{1}{\pi/2}\cdot\frac{13-6\sqrt2}{20} =\frac{13-6\sqrt2}{10\pi} }.

(d) Existence of a point attaining the average

The first branch is continuous on [−π4,0]\left[-\frac{\pi}{4},0\right]. At the endpoints of this subinterval,

f(−π4)=cos⁡(−3π4)sin⁡(−π2)=22,f(0)=0.f\left(-\frac{\pi}{4}\right) =\cos\left(-\frac{3\pi}{4}\right)\sin\left(-\frac{\pi}{2}\right) =\frac{\sqrt2}{2}, \qquad f(0)=0.

Also,

0<13−6210π<22.0<\frac{13-6\sqrt2}{10\pi}<\frac{\sqrt2}{2}.

Thus fˉ\bar f lies strictly between f(0)f(0) and f(−π4)f\left(-\frac{\pi}{4}\right). The Intermediate Value Theorem applied to the first branch gives a point

c∈(−π4,0)⊂(−π4,π4)c\in\left(-\frac{\pi}{4},0\right) \subset\left(-\frac{\pi}{4},\frac{\pi}{4}\right)

such that f(c)=fˉf(c)=\bar f. Therefore the answer is Yes.

Which branch owns the single point x=0x=0 does not affect the integral or the average. What matters is which formula is integrated on each side.

[DEFINITION]

Reduction formula

A reduction formula expresses an integral InI_n in terms of In−kI_{n-k} for a smaller index, typically by integration by parts.

[THEOREM]

Partial fractions decomposition

For a fixed denominator after cancellation and factorisation, every proper rational function ff has a unique partial-fraction expansion with terms of the form

A(x−a)k,Bx+C(x2+bx+c)k,\frac{A}{(x-a)^k},\qquad \frac{Bx+C}{(x^2+bx+c)^k},

where the quadratic x2+bx+cx^2+bx+c is irreducible over the real numbers (i.e. b2−4c<0b^2-4c<0).

[EXAMPLE]

Reduction formulae and substitution of a hyperbolic function

Consider

In=∫01(7+12x)ne7x+6x2 dx.I_n=\int_0^1(7+12x)^n e^{7x+6x^2}\,dx.

It can be written as a reduction formula in the form

In=an+bnIn−2,n≥2,I_n=a_n+b_nI_{n-2},\qquad n\ge2,

where ana_n and bnb_n depend on nn. Find ana_n and bnb_n in terms of nn.

Next, consider

J=∫cosh⁡(6x)sinh⁡4(6x)−343sinh⁡(6x) dx.J=\int\frac{\cosh(6x)} {\sinh^4(6x)-343\sinh(6x)}\,dx.

Suggest a substitution so that

J=16∫dtt4−343t,J=\frac16\int\frac{dt}{t^4-343t},

and then write the rational integrand as

A(t)t+B(t)t−7+C(t)t2+7t+49, \frac{A(t)}{t}+\frac{B(t)}{t-7} +\frac{C(t)}{t^2+7t+49},

where the degree of each numerator is less than the degree of its denominator. Find A(t)A(t), B(t)B(t) and C(t)C(t).

Solution

In the first integral, the factor 7+12x7+12x is the derivative of the exponent:

ddx(7x+6x2)=7+12x.\frac{d}{dx}(7x+6x^2)=7+12x.

Differentiating a product with one lower power of 7+12x7+12x therefore produces both an InI_n term and an In−2I_{n-2} term.

(a) Finding the reduction formula

For n≥2n\ge2, differentiate

(7+12x)n−1e7x+6x2.(7+12x)^{n-1}e^{7x+6x^2}.

Using the product rule and the chain rule gives

ddx((7+12x)n−1e7x+6x2)=12(n−1)(7+12x)n−2e7x+6x2+(7+12x)ne7x+6x2.\begin{aligned} \frac{d}{dx}\left((7+12x)^{n-1}e^{7x+6x^2}\right) &=12(n-1)(7+12x)^{n-2}e^{7x+6x^2}\\ &\quad +(7+12x)^{n}e^{7x+6x^2}. \end{aligned}

Integrate from 00 to 11. The second term on the right is InI_n, while the first term is 12(n−1)In−212(n-1)I_{n-2}, so

[(7+12x)n−1e7x+6x2]01=12(n−1)In−2+In,In=[(7+12x)n−1e7x+6x2]01−12(n−1)In−2.\begin{aligned} \left[(7+12x)^{n-1}e^{7x+6x^2}\right]_0^1 &=12(n-1)I_{n-2}+I_n,\\ I_n &=\left[(7+12x)^{n-1}e^{7x+6x^2}\right]_0^1 -12(n-1)I_{n-2}. \end{aligned}

At the endpoints,

7+12(1)=19,7+12(0)=7,7+12(1)=19,\qquad 7+12(0)=7,

and

7(1)+6(1)2=13,7(0)+6(0)2=0.7(1)+6(1)^2=13,\qquad 7(0)+6(0)^2=0.

Therefore

In=(19n−1e13−7n−1)−12(n−1)In−2.I_n =\left(19^{n-1}e^{13}-7^{n-1}\right)-12(n-1)I_{n-2}.

Comparing this with In=an+bnIn−2I_n=a_n+b_nI_{n-2} gives

an=19n−1e13−7n−1,bn=−12(n−1).{a_n=19^{n-1}e^{13}-7^{n-1}}, \qquad {b_n=-12(n-1)}.

(b) Substitution and partial fractions

The numerator cosh⁡(6x)\cosh(6x) is the derivative of sinh⁡(6x)\sinh(6x) up to a factor of 66, so this is an ordinary substitution, not a surd-clearing hyperbolic substitution. Set

t=sinh⁡(6x),dt=6cosh⁡(6x) dx.t=\sinh(6x), \qquad dt=6\cosh(6x)\,dx.

Thus cosh⁡(6x) dx=16 dt\cosh(6x)\,dx=\frac16\,dt and

J=16∫dtt4−343t.J =\frac16\int\frac{dt}{t^4-343t}.

Factor the denominator:

t4−343t=t(t3−343)=t(t−7)(t2+7t+49).t^4-343t =t(t^3-343) =t(t-7)(t^2+7t+49).

The quadratic has discriminant

72−4(49)=−147<0,7^2-4(49)=-147<0,

so it is an irreducible real quadratic. The required partial-fraction form is

1t(t−7)(t2+7t+49)=At+Bt−7+Ct+Dt2+7t+49.\frac{1}{t(t-7)(t^2+7t+49)} =\frac{A}{t}+\frac{B}{t-7} +\frac{Ct+D}{t^2+7t+49}.

The question writes the last numerator as C(t)C(t); here C(t)=Ct+DC(t)=Ct+D.

Multiplying through by the denominator gives the polynomial identity

1=A(t−7)(t2+7t+49)+Bt(t2+7t+49)+(Ct+D)t(t−7).1=A(t-7)(t^2+7t+49) +Bt(t^2+7t+49) +(Ct+D)t(t-7).

Since

(t−7)(t2+7t+49)=t3−343,(t-7)(t^2+7t+49)=t^3-343,

substitution of t=0t=0 gives

1=A(−7)(49)=−343A,A=−1343.1=A(-7)(49)=-343A, \qquad A=-\frac1{343}.

Substitution of t=7t=7 gives

1=B(7)(72+7⋅7+49)=B(7)(147)=1029B,B=11029.1=B(7)(7^2+7\cdot7+49) =B(7)(147)=1029B, \qquad B=\frac1{1029}.

To find the quadratic numerator, expand the identity and compare the coefficients of tt and t2t^2:

1=A(t3−343)+B(t3+7t2+49t)+Ct3+(D−7C)t2−7Dt.1=A(t^3-343)+B(t^3+7t^2+49t) +C t^3+(D-7C)t^2-7Dt.

The coefficient of tt gives

49B−7D=0⟹D=7B=71029.49B-7D=0 \quad\Longrightarrow\quad D=7B=\frac7{1029}.

The coefficient of t2t^2 then gives

7B+D−7C=0⟹C=7B+D7=21029.7B+D-7C=0 \quad\Longrightarrow\quad C=\frac{7B+D}{7} =\frac2{1029}.

Hence the requested numerators are

A(t)=−1343,B(t)=11029,C(t)=2t+71029.{A(t)=-\frac1{343}}, \qquad {B(t)=\frac1{1029}}, \qquad {C(t)=\frac{2t+7}{1029}}.

As a check, the decomposition is

1t4−343t=−1343t+11029(t−7)+2t+71029(t2+7t+49).\frac{1}{t^4-343t} = -\frac{1}{343t} +\frac{1}{1029(t-7)} +\frac{2t+7}{1029(t^2+7t+49)}.

Consequently, on any interval that does not contain a point where sinh⁡(6x)=0\sinh(6x)=0 or sinh⁡(6x)=7\sinh(6x)=7, an antiderivative is

J=−12058ln⁡∣t∣+16174ln⁡∣t−7∣+16174ln⁡(t2+7t+49)+C,t=sinh⁡(6x).J =-\frac1{2058}\ln|t| +\frac1{6174}\ln|t-7| +\frac1{6174}\ln(t^2+7t+49)+C, \qquad t=\sinh(6x).

The first part is a reduction formula: differentiating a product with one lower power gives a recurrence that drops the index by two. The second is a substitution followed by partial fractions for a proper rational function.

On this page