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 double Riemann sums, integrability, upper sums and changes of variables in elliptical regions.

These past exam questions test the Riemann-sum definition of integrability and an upper sum, then a change of variables with its Jacobian on an elliptical region.

[DEFINITION]

Double Riemann sum

For a partition of RR into rectangles RijR_{ij} and tag points (xij∗,yij∗)∈Rij(x_{ij}^*,y_{ij}^*)\in R_{ij}, the corresponding double Riemann sum is

∑i=1m∑j=1nf(xij∗,yij∗) ΔAij.\sum_{i=1}^{m}\sum_{j=1}^{n} f(x_{ij}^*,y_{ij}^*)\,\Delta A_{ij}.
[DEFINITION]

Upper and lower double Riemann sums

For a partition PP of RR into rectangles RijR_{ij}, the lower and upper double Riemann sums are

L(f,P)=∑i,jmij ΔAij,U(f,P)=∑i,jMij ΔAij,L(f,P)=\sum_{i,j}m_{ij}\,\Delta A_{ij}, \qquad U(f,P)=\sum_{i,j}M_{ij}\,\Delta A_{ij},

where mijm_{ij} and MijM_{ij} are the infimum and supremum of ff on RijR_{ij}.

[EXAMPLE]

Riemann integrability and an upper sum

Let f:R2→Rf:\mathbb R^2\to\mathbb R be a function defined on a rectangular region D⊂R2D\subset\mathbb R^2. Let PP be any partition of DD into subrectangles {Rij}\{R_{ij}\}, let (xij∗,yij∗)(x_{ij}^*,y_{ij}^*) be points in RijR_{ij}, and let L(f,P)L(f,P) and U(f,P)U(f,P) denote the lower and upper sums of ff over PP. Assume that ff is bounded on DD.

(a)

Which of the following guarantees that ff is integrable over DD? Select all answers that apply.

  1. inf⁡PL(f,P)=sup⁡PU(f,P)\displaystyle \inf_P L(f,P)=\sup_P U(f,P).
  2. For each ε>0\varepsilon>0 there exists a partition PP of DD such that L(f,P)−U(f,P)<ε\displaystyle L(f,P)-U(f,P)<\varepsilon.
  3. For each ε>0\varepsilon>0 there exists a partition PP of DD such that U(f,P)−L(f,P)<ε\displaystyle U(f,P)-L(f,P)<\varepsilon.
  4. For every sequence of tagged rectangular partitions (Pm)(P_m) of DD whose mesh tends to zero, the corresponding tagged sums converge to a common limit independently of the choice of tags.
  5. f(x,y)≡0f(x,y)\equiv0 on R2∖D\mathbb R^2\setminus D.

(b)

Kenneth is considering the problem of approximating the double integral of a function using an upper Riemann sum. Let

R1=[3,9]×[2,7],f:R2→R,f(x,y)=7x2y2.R_1=[3,9]\times[2,7], \qquad f:\mathbb R^2\to\mathbb R, \qquad f(x,y)=7x^2y^2.

Let PxP_x be a partition of [3,9][3,9] into 33 equal parts and let PyP_y be a partition of [2,7][2,7] into 22 equal parts. Let PP be the partition generated by the rectangular regions associated to Px×PyP_x\times P_y.

Find the area AA of each subrectangle and compute the upper Riemann sum U(f,P)U(f,P). Identify the corresponding boxes relative to the graph of ff.

Solution

Part (a) tests two equivalent forms of the definition of integrability: upper and lower sums bound ff on each subrectangle, while tagged sums use arbitrary sample points. Part (b) uses the fact that ff is increasing on the rectangle, so its maximum on each subrectangle is at the upper-right corner.

(a) Selecting the integrability criteria

For a partition PP, the lower and upper sums satisfy

L(f,P)≤U(f,P).L(f,P)\le U(f,P).

The upper/lower-sum criterion says that ff is integrable when, for every ε>0\varepsilon>0, there is a partition for which

0≤U(f,P)−L(f,P)<ε.0\le U(f,P)-L(f,P)<\varepsilon.

Therefore statement 3 is correct.

Statement 1 is a strong sufficient condition. If c=inf⁡PL(f,P)=sup⁡PU(f,P)c=\inf_P L(f,P)=\sup_P U(f,P), then for every partition PP,

inf⁡PL(f,P)≤L(f,P)≤U(f,P)≤sup⁡PU(f,P),\inf_P L(f,P)\le L(f,P)\le U(f,P)\le\sup_P U(f,P),

so L(f,P)=U(f,P)=cL(f,P)=U(f,P)=c. The gap between upper and lower sums is zero, so ff is integrable. The condition is stronger than necessary, but it does guarantee integrability.

Statement 2 has the wrong sign: since L(f,P)−U(f,P)≤0L(f,P)-U(f,P)\le0 always, the inequality says nothing about the size of the gap. Statement 5 concerns values outside DD, whereas integrability over DD depends only on ff on DD.

Statement 4 is the tagged-sum form of the definition: the partitions become arbitrarily fine and the limit does not depend on the tags. Without that independence, different choices of tags could give different limits. Hence the correct selections are

statements 1, 3 and 4.{\text{statements 1, 3 and 4}.}

(b) Computing the upper sum

The width of R1R_1 is 9−3=69-3=6. Dividing it into 33 equal parts gives

Δx=63=2.\Delta x=\frac63=2.

The height is 7−2=57-2=5. Dividing it into 22 equal parts gives

Δy=52.\Delta y=\frac52.

Thus the area of every subrectangle is

A=Δx Δy=2⋅52=5.{A=\Delta x\,\Delta y =2\cdot\frac52=5}.

The partition points are

Px={3,5,7,9},Py={2,92,7}.P_x=\{3,5,7,9\}, \qquad P_y=\left\{2,\frac92,7\right\}.

On R1R_1, both xx and yy are positive and

∂f∂x=14xy2>0,∂f∂y=14x2y>0.\frac{\partial f}{\partial x}=14xy^2>0, \qquad \frac{\partial f}{\partial y}=14x^2y>0.

So ff is increasing in both variables and its maximum on each subrectangle is at the upper-right corner. The upper-right xx-values are 5,7,95,7,9 and the upper-right yy-values are 92,7\frac92,7. Hence

U(f,P)=5[7(52)(92)2+7(52)(72)+7(72)(92)2+7(72)(72)+7(92)(92)2+7(92)(72)]=5⋅7(52+72+92)((92)2+72)=35(155)(814+49)=35(155)(2774)=15027254.\begin{aligned} U(f,P) &=5\left[ 7(5^2)\left(\frac92\right)^2 +7(5^2)(7^2) +7(7^2)\left(\frac92\right)^2 +7(7^2)(7^2)\right.\\ &\qquad\left. +7(9^2)\left(\frac92\right)^2 +7(9^2)(7^2) \right]\\ &=5\cdot7 \left(5^2+7^2+9^2\right) \left(\left(\frac92\right)^2+7^2\right)\\ &=35(155)\left(\frac{81}{4}+49\right)\\ &=35(155)\left(\frac{277}{4}\right)\\ &={\frac{1502725}{4}}. \end{aligned}

This upper sum is the total volume of the boxes on the subrectangles of PP that reach above the graph. The lower sum would use the lower-left corners, giving boxes that fit below the graph, as in the double Riemann sum definition.

[THEOREM]

Polar area element

For r≥0r\ge0, the polar-coordinate change of variables is

x=rcos⁡θ,y=rsin⁡θ,x=r\cos\theta,\qquad y=r\sin\theta,

and its Jacobian satisfies

∣∂(x,y)∂(r,θ)∣=r.\left|\frac{\partial(x,y)}{\partial(r,\theta)}\right| =r.

Consequently,

dA=r dr dθ,dA=r\,dr\,d\theta,

and

∬Ωf(x,y) dA=∬Φf(rcos⁡θ,rsin⁡θ) r dr dθ.\iint_\Omega f(x,y)\,dA =\iint_\Phi f(r\cos\theta,r\sin\theta)\,r\,dr\,d\theta.
[EXAMPLE]

Changing variables in an elliptical integral

Your friend Brad is interested in calculating the following integral:

J=∬Ω(3x+2y)2 dx dy.(1)J=\iint_{\Omega}(3x+2y)^2\,dx\,dy. \tag{1}

Here Ω\Omega is the region bounded by

y=0andy=5442−x2.y=0 \qquad\text{and}\qquad y=\frac54\sqrt{4^2-x^2}.

(a) Choosing the change of variables

Brad wants to find a change of variables to write this integral in terms of coordinates that are very similar to the standard polar coordinates. Choose the suitable change of variables for the region Ω\Omega from the following options:

  1. x=4rcos⁡θ, y=5rsin⁡θx=4r\cos\theta,\ y=5r\sin\theta where 0≤θ≤2π0\le\theta\le2\pi and −1≤r≤1-1\le r\le1.
  2. x=4rcos⁡θ, y=5rsin⁡θx=4r\cos\theta,\ y=5r\sin\theta where 0≤θ≤π0\le\theta\le\pi and −1≤r≤1-1\le r\le1.
  3. x=4rcos⁡θ, y=5rsin⁡θx=4r\cos\theta,\ y=5r\sin\theta where −π≤θ≤0-\pi\le\theta\le0 and 0≤r≤10\le r\le1.
  4. x=4rcos⁡θ, y=5rsin⁡θx=4r\cos\theta,\ y=5r\sin\theta where 0≤θ≤2π0\le\theta\le2\pi and 0≤r≤10\le r\le1.
  5. x=4rcos⁡θ, y=5rsin⁡θx=4r\cos\theta,\ y=5r\sin\theta where 0≤θ≤π0\le\theta\le\pi and 0≤r≤10\le r\le1.
  6. x=4rcos⁡θ, y=5rsin⁡θx=4r\cos\theta,\ y=5r\sin\theta where −π≤θ≤0-\pi\le\theta\le0 and −1≤r≤1-1\le r\le1.

The suitable option is 5:

x=4rcos⁡θ,y=5rsin⁡θ,0≤θ≤π,0≤r≤1.x=4r\cos\theta,\qquad y=5r\sin\theta, \qquad 0\le\theta\le\pi,\qquad 0\le r\le1.

To check the region, substitute the transformation into the equation of the curved boundary:

x216+y225=r2cos⁡2θ+r2sin⁡2θ=r2.\frac{x^2}{16}+\frac{y^2}{25} =r^2\cos^2\theta+r^2\sin^2\theta =r^2.

Thus r=1r=1 gives the ellipse

x216+y225=1,\frac{x^2}{16}+\frac{y^2}{25}=1,

whose upper half is the given curve. The condition 0≤r≤10\le r\le1 fills the inside of the ellipse and 0≤θ≤π0\le\theta\le\pi gives y=5rsin⁡θ≥0y=5r\sin\theta\ge0.

The other angular ranges describe the lower half or the full ellipse. Allowing negative rr gives some points a second representation.

(b) Finding g(r,θ)g(r,\theta)

With the order of integration requested in the question, the transformed integral has the form

J=∫a1a2∫θ1θ2g(r,θ) dθ dr.(2)J=\int_{a_1}^{a_2}\int_{\theta_1}^{\theta_2} g(r,\theta)\,d\theta\,dr. \tag{2}

The Jacobian matrix is

∂(x,y)∂(r,θ)=(4cos⁡θ−4rsin⁡θ5sin⁡θ5rcos⁡θ).\frac{\partial(x,y)}{\partial(r,\theta)} = \begin{pmatrix} 4\cos\theta&-4r\sin\theta\\ 5\sin\theta&5r\cos\theta \end{pmatrix}.

Therefore,

det⁡∂(x,y)∂(r,θ)=det⁡(4cos⁡θ−4rsin⁡θ5sin⁡θ5rcos⁡θ)=20rcos⁡2θ+20rsin⁡2θ=20r.\begin{aligned} \det\frac{\partial(x,y)}{\partial(r,\theta)} &= \det \begin{pmatrix} 4\cos\theta&-4r\sin\theta\\ 5\sin\theta&5r\cos\theta \end{pmatrix}\\ &=20r\cos^2\theta+20r\sin^2\theta\\ &=20r. \end{aligned}

Since 0≤r≤10\le r\le1, the absolute Jacobian factor is 20r20r. The integrand becomes

(3x+2y)2=(12rcos⁡θ+10rsin⁡θ)2=r2(12cos⁡θ+10sin⁡θ)2.(3x+2y)^2 =\bigl(12r\cos\theta+10r\sin\theta\bigr)^2 =r^2\bigl(12\cos\theta+10\sin\theta\bigr)^2.

Multiplying by the area factor gives

g(r,θ)=20r3(12cos⁡θ+10sin⁡θ)2.{ g(r,\theta) =20r^3\bigl(12\cos\theta+10\sin\theta\bigr)^2 }.

Here a1=0a_1=0, a2=1a_2=1, θ1=0\theta_1=0 and θ2=π\theta_2=\pi.

(c) Calculating JJ

Substituting the transformed integrand and bounds into (2) gives

J=∫01∫0π20r3(12cos⁡θ+10sin⁡θ)2 dθ dr.J=\int_0^1\int_0^\pi 20r^3\bigl(12\cos\theta+10\sin\theta\bigr)^2 \,d\theta\,dr.

The radial factor is

∫0120r3 dr=20[r44]01=5.\int_0^1 20r^3\,dr =20\left[\frac{r^4}{4}\right]_0^1 =5.

For the angular factor, expand the square:

∫0π(12cos⁡θ+10sin⁡θ)2 dθ=∫0π(144cos⁡2θ+240sin⁡θcos⁡θ+100sin⁡2θ) dθ=144⋅π2+240[sin⁡2θ2]0π+100⋅π2=72π+0+50π=122π.\begin{aligned} \int_0^\pi \bigl(12\cos\theta+10\sin\theta\bigr)^2\,d\theta &=\int_0^\pi \left(144\cos^2\theta +240\sin\theta\cos\theta +100\sin^2\theta\right)\,d\theta\\ &=144\cdot\frac{\pi}{2} +240\left[\frac{\sin^2\theta}{2}\right]_0^\pi +100\cdot\frac{\pi}{2}\\ &=72\pi+0+50\pi\\ &=122\pi. \end{aligned}

Hence

J=5(122π)=610π.{J=5(122\pi)=610\pi}.

Scaled polar coordinates suit this problem because they turn the upper half ellipse into the rectangle 0≤r≤10\le r\le1, 0≤θ≤π0\le\theta\le\pi, with the Jacobian supplying the area scaling.

The vanishing mixed term reflects the symmetry of the upper half ellipse under x↦−xx\mapsto-x: the sin⁡θcos⁡θ\sin\theta\cos\theta term comes from the part of the integrand that cancels between the left and right halves.

On this page