Double Integrals Over Rectangles
Geometry
We have z=f(x,y) defined on
[
a
,
b
]
×
[
c
,
d
]
[a,b] \times [c,d]
[a,b]×[c,d]. We name the domain to be D, and cut it into smaller rectangles. Take an arbitrary point
(
x
i
j
,
y
i
j
)
(x_{ij}, y_{ij})
(xij,yij) on the rectangle
[
x
i
,
x
i
+
1
]
×
[
y
j
,
y
j
+
1
]
[x_i, x_{i+1}] \times [y_j,y_{j+1}]
[xi,xi+1]×[yj,yj+1]. We calculate the Riemann sum.
∑
i
=
1
m
∑
j
=
1
n
f
(
x
i
j
,
y
i
j
)
(
x
i
+
1
−
x
i
)
(
y
j
+
1
−
y
j
)
=
∑
i
=
1
m
∑
j
=
1
n
f
(
x
i
j
,
y
i
j
)
b
−
a
m
d
−
c
n
\sum_{i=1}^m\sum_{j=1}^n f(x_{ij}, y_{ij}) (x_{i+1} - x_i)(y_{j+1} - y_{j}) \\\\ = \sum_{i=1}^m\sum_{j=1}^n f(x_{ij}, y_{ij}) \frac{b-a}{m} \frac{d-c}{n} \\\\
i=1∑mj=1∑nf(xij,yij)(xi+1−xi)(yj+1−yj)=i=1∑mj=1∑nf(xij,yij)mb−and−c
If such a limit exists, then we define it to be the double integral of f over the rectangle D. This is the volume between f and xy-plane on D. If the value of the function is always 1 in D, then the double integral equals to the area of D.
Iterated Integrals
Introduce a function A ( x ) = ∫ c d f ( x , y ) d y A(x) = \int_c^d f(x,y) dy A(x)=∫cdf(x,y)dy. Next, we integrate A. ∫ a b A ( x ) d x = ∫ a b [ ∫ c d f ( x , y ) d y ] d x \int_a^b A(x) dx = \int_a^b [\int_c^d f(x,y) dy] dx ∫abA(x)dx=∫ab[∫cdf(x,y)dy]dx. Alternatively, we can introduce a function of B ( y ) B(y) B(y) to integrate on x first.
Fubini’s Theorem
Assume f is continuous function on the rectangle
D
=
{
(
x
,
y
)
∣
a
≤
x
≤
b
,
c
≤
y
≤
d
}
D=\{(x,y)|a\le x\le b, c \le y\le d \}
D={(x,y)∣a≤x≤b,c≤y≤d}.
∫
∫
D
f
(
x
,
y
)
d
A
=
∫
a
b
∫
c
d
f
(
x
,
y
)
d
y
d
x
=
∫
c
d
∫
a
b
f
(
x
,
y
)
d
x
d
y
\int\int_D f(x,y) dA = \int_a^b \int_c^d f(x,y) dydx = \int_c^d \int_a^b f(x,y) dxdy
∫∫Df(x,y)dA=∫ab∫cdf(x,y)dydx=∫cd∫abf(x,y)dxdy
E.x.
Evaluate ∫ ∫ y s i n ( x y ) d A \int \int ysin(xy) dA ∫∫ysin(xy)dA where D = [ 1 , 2 ] × [ 0 , π ] D = [1,2] \times [0,\pi] D=[1,2]×[0,π].
sol:
∫
∫
y
s
i
n
(
x
y
)
d
A
=
∫
0
π
∫
1
2
y
s
i
n
(
x
y
)
d
x
d
y
=
∫
0
π
[
−
c
o
s
(
x
y
)
]
1
2
d
y
=
∫
0
π
c
o
s
(
y
)
−
c
o
s
(
2
y
)
d
y
\int \int ysin(xy) dA \\\\ = \int_0^{\pi} \int_1^2 ysin(xy) dxdy \\\\ = \int_0^{\pi} [-cos(xy)]_1^2 dy \\\\ = \int_0^{\pi} cos(y)-cos(2y) dy \\\\
∫∫ysin(xy)dA=∫0π∫12ysin(xy)dxdy=∫0π[−cos(xy)]12dy=∫0πcos(y)−cos(2y)dy
The order of doing iterated integral may be crucial.
Double Integrals Over General Regions
We have f(x,y) on D. Assume D is bounded (we can find a rectangle R that encloses D).
Define h(x,y) = f(x,y) if
(
x
,
y
)
∈
D
(x,y) \in D
(x,y)∈D, and h(x,y) = 0 if
(
x
,
y
)
∈
(
R
−
D
)
(x,y) \in (R-D)
(x,y)∈(R−D). In this case,
∫
∫
D
f
(
x
,
y
)
d
A
=
∫
∫
R
h
(
x
,
y
)
d
A
\int\int_D f(x,y) dA = \int\int_R h(x,y) dA
∫∫Df(x,y)dA=∫∫Rh(x,y)dA
A plane region D is said to be of type I if it lies between the graphs of two continuous functions of x, say
g
1
(
x
)
g_1(x)
g1(x) and
g
2
(
x
)
g_2(x)
g2(x).
By applying the Fubini’s theorem,
∫
∫
D
f
(
x
,
y
)
d
A
=
∫
a
b
∫
g
1
(
x
)
g
2
(
x
)
f
(
x
,
y
)
d
y
d
x
\int\int_D f(x,y) dA = \int_a^b \int_{g_1(x)}^{g_2(x)} f(x,y) dydx
∫∫Df(x,y)dA=∫ab∫g1(x)g2(x)f(x,y)dydx
A plane region D is said to be of type II if it lies between the graphs of two continuous functions of y.
E.x.
Evaluate ∫ 0 1 ∫ x 1 s i n ( y 2 ) d y d x \int_0^1 \int_x^1 sin(y^2) dydx ∫01∫x1sin(y2)dydx.
sol:
Trick here is to switch the order of integration. Drawing the region D, we find out that it is both type I and type II. Convert the integral.
∫
0
1
∫
x
1
s
i
n
(
y
2
)
d
y
d
x
=
∫
0
1
∫
0
y
s
i
n
(
y
2
)
d
x
d
y
=
∫
0
1
y
s
i
n
(
y
2
)
d
y
\int_0^1 \int_x^1 sin(y^2) dydx \\\\ = \int_0^1 \int_0^y sin(y^2) dxdy \\\\ = \int_0^1 ysin(y^2) dy \\\\
∫01∫x1sin(y2)dydx=∫01∫0ysin(y2)dxdy=∫01ysin(y2)dy
Double Integrals In Polar Coordinates
Suppose f is a continuous function defined on D (Draw a picture of D!). Convert the limit of x and y to r and theta in the polar coordinate as
0
≤
a
≤
r
≤
b
0 \le a \le r \le b
0≤a≤r≤b,
α
≤
β
\alpha \le \beta
α≤β. The formula is as below
∫
∫
D
f
(
x
,
y
)
d
A
=
∫
α
β
∫
a
b
f
(
r
c
o
s
θ
,
r
s
i
n
θ
)
r
d
r
d
θ
\int\int_D f(x,y) dA = \int_{\alpha}^{\beta} \int_{a}^{b} f(r cos \theta, r sin \theta)r dr d\theta
∫∫Df(x,y)dA=∫αβ∫abf(rcosθ,rsinθ)rdrdθ
If the formula involves z, then simplify it to the form of z=f(x,y). If the original equation involves terms like
z
2
z^2
z2, remember the integral of
z
=
g
(
x
,
y
)
z=\sqrt{g(x,y)}
z=g(x,y) only calculate the part where
z
>
0
z>0
z>0. Based on symmetry, we can get the desired answer by multiplying the integral result by 2.
E.x.
Evaluate ∫ ∫ D 3 x + 4 y 2 d A \int \int_D 3x+4y^2 dA ∫∫D3x+4y2dA where the region is the upper half plane bounded by the circles x 2 + y 2 = 1 x^2+y^2=1 x2+y2=1 and x 2 + y 2 = 4 x^2+y^2=4 x2+y2=4.
sol:
Write
D
=
{
(
r
,
θ
)
∣
1
≤
r
≤
2
,
0
≤
θ
≤
π
}
D = \{(r,\theta) | 1 \le r \le 2, 0 \le \theta \le \pi\}
D={(r,θ)∣1≤r≤2,0≤θ≤π}.
∫
∫
D
3
x
+
4
y
2
d
A
=
∫
1
2
∫
0
π
(
3
r
c
o
s
θ
+
4
r
2
s
i
n
2
θ
)
r
d
θ
d
r
\int \int_D 3x+4y^2 dA = \int_1^2 \int_0^{\pi} (3rcos\theta + 4r^2sin^2\theta)r d\theta dr
∫∫D3x+4y2dA=∫12∫0π(3rcosθ+4r2sin2θ)rdθdr
E.x.
Compute the volume of the solid bounded by the plane z=0 and the paraboloid z = 1 − x 2 − y 2 z=1-x^2-y^2 z=1−x2−y2.
sol:
∫
∫
D
1
−
x
2
−
y
2
d
A
=
∫
0
1
∫
0
2
π
(
1
−
r
2
)
r
d
θ
d
r
\int \int_D 1-x^2-y^2 dA = \int_0^1 \int_0^{2\pi} (1-r^2)r d\theta dr
∫∫D1−x2−y2dA=∫01∫02π(1−r2)rdθdr
E.x.
Find the region enclosed by r = c o s 2 θ r=cos2\theta r=cos2θ. For simplicity, restrict − π / 4 ≤ θ ≤ π / 4 -\pi/4 \le \theta \le \pi/4 −π/4≤θ≤π/4.
sol:
a
r
e
a
=
∫
π
/
4
−
π
/
4
∫
0
c
o
s
2
θ
r
d
r
d
θ
=
∫
π
/
4
−
π
/
4
1
2
(
c
o
s
2
θ
)
2
d
θ
=
∫
π
/
4
−
π
/
4
(
1
+
c
o
s
4
θ
)
4
d
θ
=
π
/
8
area = \int_{\pi/4}^{-\pi/4} \int_{0}^{cos 2 \theta} r dr d\theta\\\\ = \int_{\pi/4}^{-\pi/4} \frac 1 2 (cos 2\theta)^2 d\theta \\\\ = \int_{\pi/4}^{-\pi/4} \frac{(1 + cos 4\theta)}{4} d\theta \\\\ = \pi/8
area=∫π/4−π/4∫0cos2θrdrdθ=∫π/4−π/421(cos2θ)2dθ=∫π/4−π/44(1+cos4θ)dθ=π/8
Applications of Double Integrals
Surface Area Element
S is a surface given by z=f(x,y). Derive a formula for computing the area of S. Assume that the domain of f is rectangular. The idea is to approximate S by tangent planes. Let Δ T i j \Delta T_{ij} ΔTij to represent such planes. The work remaining is to compute Δ T i j \Delta T_{ij} ΔTij.
Name the points of one such plane as P, Q, R, W. The vector
P
Q
⃗
=
<
Δ
x
,
0
,
f
x
(
x
i
,
y
j
)
Δ
x
>
\vec{PQ} = <\Delta x, 0, f_x(x_i,y_j)\Delta x>
PQ=<Δx,0,fx(xi,yj)Δx> , and the vector
P
R
⃗
=
<
0
,
Δ
y
,
f
y
(
x
i
,
y
j
)
Δ
y
>
\vec{PR} = <0, \Delta y, f_y(x_i,y_j)\Delta y>
PR=<0,Δy,fy(xi,yj)Δy> . The area of the parallelogram PQWR is given by
∣
P
Q
⃗
×
P
R
⃗
∣
=
Δ
x
Δ
y
∣
<
−
f
x
(
x
i
,
y
j
)
,
−
f
y
(
x
i
,
y
j
)
,
1
>
∣
|\vec{PQ} \times \vec{PR}| = \Delta x\Delta y|<-f_x(x_i,y_j),-f_y(x_i,y_j),1>|
∣PQ×PR∣=ΔxΔy∣<−fx(xi,yj),−fy(xi,yj),1>∣
Using Riemann sum, we obtain the area of S as follows
a
r
e
a
(
S
)
=
∫
∫
D
f
x
2
+
f
y
2
+
1
d
x
d
y
area(S) = \int \int_D \sqrt{f_x^2 + f_y^2 + 1} dxdy
area(S)=∫∫Dfx2+fy2+1dxdy
Density and Mass
The density at the point (x,y) is given by
ρ
(
x
,
y
)
=
l
i
m
Δ
m
Δ
A
\rho(x,y) = lim \frac{\Delta m}{\Delta A}
ρ(x,y)=limΔAΔm
Assume we know that density
ρ
(
x
,
y
)
\rho(x,y)
ρ(x,y) and we would like to compute the mass of the object.
m
=
∫
∫
D
ρ
(
x
,
y
)
d
A
m = \int \int_D \rho(x,y) dA
m=∫∫Dρ(x,y)dA
The same idea can be applied to the charge density and total charge.
Center of Mass
The moment of the lamina about the x-axis and y-axis is
M
x
=
∫
∫
D
y
ρ
(
x
,
y
)
d
A
M
y
=
∫
∫
D
x
ρ
(
x
,
y
)
d
A
M_x = \int\int_D y\rho(x,y) dA \\\\ M_y = \int\int_D x\rho(x,y) dA \\\\
Mx=∫∫Dyρ(x,y)dAMy=∫∫Dxρ(x,y)dA
The center of the mass is
(
x
ˉ
,
y
ˉ
)
(\bar x, \bar y)
(xˉ,yˉ) so that
m
x
ˉ
=
M
y
m \bar x = M_y
mxˉ=My and
m
y
ˉ
=
M
x
m \bar y = M_x
myˉ=Mx. Therefore,
x
ˉ
=
1
m
∫
∫
D
x
ρ
(
x
,
y
)
d
A
y
ˉ
=
1
m
∫
∫
D
y
ρ
(
x
,
y
)
d
A
\bar x = \frac 1 m \int\int_D x\rho(x,y) dA \\\\ \bar y = \frac 1 m \int\int_D y\rho(x,y) dA \\\\
xˉ=m1∫∫Dxρ(x,y)dAyˉ=m1∫∫Dyρ(x,y)dA
E.x.
The region is a half-circle with center at the origin and radius of a. The density of a lamina is x 2 + y 2 \sqrt{x^2+y^2} x2+y2. Find its center of mass.
sol:
Based on symmetry, we know
M
x
M_x
Mx is 0.
M
y
=
∫
∫
D
y
x
2
+
y
2
d
A
=
∫
0
π
∫
0
a
r
3
s
i
n
θ
d
r
d
θ
M_y = \int\int_D y\sqrt{x^2+y^2} dA \\\\ = \int_0^{\pi}\int_0^a r^3 sin\theta \ drd\theta
My=∫∫Dyx2+y2dA=∫0π∫0ar3sinθ drdθ
Probability
Let f(x,y) stand for the pmf.
P
[
(
x
,
y
)
∈
D
]
=
∫
∫
D
f
(
x
,
y
)
d
A
P[(x,y) \in D] = \int\int_D f(x,y) dA
P[(x,y)∈D]=∫∫Df(x,y)dA
Triple Integrals
The Fubini’s Theorem can be extended to triple integrals.
E in
R
3
R^3
R3 is said to be of type I if it can be written as
E
=
{
(
x
,
y
,
z
)
∣
(
x
,
y
)
∈
D
,
D
⊂
R
2
,
u
1
(
x
,
y
)
≤
z
≤
u
2
(
x
,
y
)
}
∫
∫
∫
E
f
(
x
,
y
,
z
)
d
V
=
∫
∫
D
[
∫
u
1
(
x
,
y
)
u
2
(
x
,
y
)
f
(
x
,
y
,
z
)
d
z
]
d
A
E = \{(x,y,z)|(x,y) \in D, D \sub R^2, u_1(x,y) \le z \le u_2(x,y) \} \\\\ \int\int\int_E f(x,y,z) dV = \int\int_D[\int_{u_1(x,y)}^{u_2(x,y)} f(x,y,z) dz] dA
E={(x,y,z)∣(x,y)∈D,D⊂R2,u1(x,y)≤z≤u2(x,y)}∫∫∫Ef(x,y,z)dV=∫∫D[∫u1(x,y)u2(x,y)f(x,y,z)dz]dA
If we also know that D is, say, type I, then the above triple is further equal to
∫
a
b
∫
g
1
(
x
)
g
2
(
x
)
∫
u
1
(
x
,
y
)
u
2
(
x
,
y
)
f
(
x
,
y
,
z
)
d
z
d
y
d
x
\int_a^b\int_{g_1(x)}^{g_2(x)}\int_{u_1(x,y)}^{u_2(x,y)} f(x,y,z) dzdydx
∫ab∫g1(x)g2(x)∫u1(x,y)u2(x,y)f(x,y,z)dzdydx
E.x.
E is bounded by the coordinates and the plane x+y+z=1. Find ∫ ∫ ∫ E z d V \int\int\int_E z dV ∫∫∫EzdV.
sol:
Define the projection of E to the xy-plane to be D.
E
=
{
(
x
,
y
,
z
)
∣
(
x
,
y
)
∈
D
,
0
≤
z
≤
1
−
x
−
y
}
∫
∫
∫
E
z
d
V
=
∫
0
1
∫
0
1
−
x
∫
0
1
−
x
−
y
z
d
z
d
y
d
x
E = \{(x,y,z)|(x,y) \in D, 0 \le z \le 1-x-y \} \\\\ \int\int\int_E z dV = \int_0^1\int_{0}^{1-x}\int_{0}^{1-x-y} z dzdydx
E={(x,y,z)∣(x,y)∈D,0≤z≤1−x−y}∫∫∫EzdV=∫01∫01−x∫01−x−yzdzdydx
E.x.
Use triple integrals to find the volume of the tetrahedron T bounded by the plane x+2y+z=2, x=2y, x=0, and z=0.
sol:
Draw the picture. Find the intersection. A=(0,0,2), B=(0,1,0), C=(1,1/2,0). We want to find volume of OABC.
V
O
A
B
C
=
∫
∫
∫
E
1
d
V
=
∫
0
1
∫
x
/
2
1
−
x
/
2
∫
0
2
−
x
−
2
y
d
z
d
y
d
x
V_{OABC} = \int\int\int_E 1 dV \\\\ = \int_0^1\int_{x/2}^{1-x/2}\int_{0}^{2-x-2y} dzdydx \\\\
VOABC=∫∫∫E1dV=∫01∫x/21−x/2∫02−x−2ydzdydx
Triple Integrals in Cylindrical Coordinates
Say E is a type I and D has a clean form in the polar coordinate.
D
=
{
(
r
,
θ
)
∣
α
≤
θ
≤
β
,
h
1
(
θ
)
≤
r
≤
h
2
(
θ
)
}
D = \{(r,\theta)| \alpha \le \theta \le \beta, h_1(\theta) \le r \le h_2(\theta) \}
D={(r,θ)∣α≤θ≤β,h1(θ)≤r≤h2(θ)}
It is just the polar coordinates adding the z-axis.
∫
∫
∫
E
f
(
x
,
y
,
z
)
d
V
=
∫
∫
∫
V
f
(
r
,
θ
,
z
)
r
d
z
d
r
d
θ
\int\int\int_E f(x,y,z) dV = \int\int\int_V f(r,\theta, z)r \ dz drd\theta
∫∫∫Ef(x,y,z)dV=∫∫∫Vf(r,θ,z)r dzdrdθ
E.x.
Describe the surface given by z=r in the cylindrical coordinate.
sol:
z
=
r
=
x
2
+
y
2
z = r = \sqrt{x^2+y^2}
z=r=x2+y2
E.x.
A solid E lies within the cylinder x 2 + y 2 = 1 x^2+y^2=1 x2+y2=1, below the plane z=4 and above the paraboloid z = 1 − ( x 2 + y 2 ) z=1-(x^2+y^2) z=1−(x2+y2). The density function is x 2 + y 2 \sqrt{x^2+y^2} x2+y2. Find its mass.
sol:
∫
∫
∫
E
ρ
(
x
,
y
,
z
)
d
V
=
∫
0
2
π
∫
0
1
∫
1
−
r
2
4
r
2
d
z
d
r
d
θ
\int\int\int_E \rho(x,y,z) dV \\\\ = \int_0^{2\pi}\int_0^1\int_{1-r^2}^{4} r^2 \ dzdrd\theta \\\\
∫∫∫Eρ(x,y,z)dV=∫02π∫01∫1−r24r2 dzdrdθ
Triple Integrals in Spherical Coordinates
We use
(
ρ
,
θ
,
ϕ
)
(\rho, \theta, \phi)
(ρ,θ,ϕ) to represent a point P in the spherical coordinates. The first component is the distance between O and P. The third component is the angle between
O
P
⃗
\vec{OP}
OP and the positive direction of z-axis (
o
≤
ϕ
≤
π
o \le \phi \le \pi
o≤ϕ≤π).
x
=
ρ
s
i
n
ϕ
c
o
s
θ
y
=
ρ
s
i
n
ϕ
s
i
n
θ
z
=
ρ
c
o
s
ϕ
x = \rho sin \phi cos \theta \\\\ y = \rho sin \phi sin \theta \\\\ z = \rho cos \phi \\\\
x=ρsinϕcosθy=ρsinϕsinθz=ρcosϕ
The formula is as below
∫
∫
∫
E
f
(
x
,
y
,
z
)
d
V
=
∫
c
d
∫
α
β
∫
a
b
f
(
ρ
,
θ
,
ϕ
)
ρ
2
s
i
n
ϕ
d
ρ
d
θ
d
ϕ
\int\int\int_E f(x,y,z) dV = \int_{c}^{d} \int_{\alpha}^{\beta} \int_{a}^{b} f(\rho, \theta, \phi) \rho^2 sin \phi \ d\rho d \theta d \phi
∫∫∫Ef(x,y,z)dV=∫cd∫αβ∫abf(ρ,θ,ϕ)ρ2sinϕ dρdθdϕ
E.x.
Compute the volume between the cone z = x 2 + y 2 z=\sqrt{x^2+y^2} z=x2+y2 and the sphere x 2 + y 2 + z 2 = z x^2+y^2+z^2=z x2+y2+z2=z.
sol:
The latter part is a sphere of radius 0.5 centered at (0,0,0.5). In the spherical coordinates, the two planes can be expressed as
c
o
s
ϕ
=
s
i
n
ϕ
ρ
=
c
o
s
ϕ
cos \phi = sin \phi\\\\ \rho = cos \phi \\\\
cosϕ=sinϕρ=cosϕ
The solid can be expressed as
E
=
{
(
ρ
,
θ
,
ϕ
)
∣
0
≤
ϕ
≤
π
/
4
,
0
≤
ρ
≤
c
o
s
ϕ
,
0
≤
θ
≤
2
π
}
E = \{(\rho, \theta, \phi)| 0 \le \phi \le \pi/4, 0 \le \rho \le cos \phi, 0 \le \theta \le 2\pi\}
E={(ρ,θ,ϕ)∣0≤ϕ≤π/4,0≤ρ≤cosϕ,0≤θ≤2π}
Change of Variables in Multiple Integrals
One-to-One Maps
Change of variable for integrals.
We have
f
:
[
a
,
b
]
−
>
R
f:[a,b]->R
f:[a,b]−>R and want to substitute
x
=
g
(
u
)
x=g(u)
x=g(u). Let
g
(
c
)
=
a
,
g
(
d
)
=
b
g(c)=a, g(d)=b
g(c)=a,g(d)=b. We need one assumption that g is monotonic.
∫
a
b
f
(
x
)
d
x
=
∫
c
d
f
(
g
(
u
)
)
g
′
(
u
)
d
u
\int_a^b f(x)dx = \int_c^d f(g(u))g'(u) du
∫abf(x)dx=∫cdf(g(u))g′(u)du
Change of variable for double integrals. Special case: polar coordinate.
In general, transformation
T
(
u
,
v
)
=
(
x
,
y
)
T(u,v) = (x,y)
T(u,v)=(x,y). (u,v) and (x,y) are connected by
x
=
g
(
u
,
v
)
y
=
h
(
u
,
v
)
x = g(u,v)\\\\ y = h(u,v)
x=g(u,v)y=h(u,v)
If it happens that for
(
u
1
,
v
1
)
≠
(
u
2
,
v
2
)
(u_1, v_1) \ne (u_2, v_2)
(u1,v1)=(u2,v2) and
T
(
u
1
,
v
1
)
=
T
(
u
2
,
v
2
)
T(u_1, v_1) = T(u_2, v_2)
T(u1,v1)=T(u2,v2) then we say that the map T is not one-to-one.
Given ( x 2 , y 2 ) (x_2, y_2) (x2,y2), we can find one ( u 2 , v 2 ) (u_2, v_2) (u2,v2) s.t. T ( u 2 , v 2 ) = ( x 2 , y 2 ) T(u_2, v_2)=(x_2, y_2) T(u2,v2)=(x2,y2). This is called T − 1 T^{-1} T−1.
To see whether a map is one-to-one:
-
T sends different points to different points
-
for every ( x , y ) ∈ R 2 (x,y)\in R^2 (x,y)∈R2, we are able to find ( u , v ) (u, v) (u,v) s.t. T ( u , v ) = ( x , y ) T(u, v)=(x, y) T(u,v)=(x,y).
Change Variable Formula
We approximate the region after transformation by a parallelogram.
The area of such parallelogram is
A
B
⃗
=
T
(
u
+
Δ
u
,
v
)
−
T
(
u
,
v
)
=
<
g
(
u
+
Δ
u
,
v
)
−
g
(
u
,
v
)
,
h
(
u
+
Δ
u
,
v
)
−
h
(
u
,
v
)
>
=
<
g
u
Δ
u
,
h
u
Δ
u
>
A
C
⃗
=
T
(
u
,
v
+
Δ
v
)
−
T
(
u
,
v
)
=
<
g
(
u
,
v
+
Δ
v
)
−
g
(
u
,
v
)
,
h
(
u
,
v
+
Δ
v
)
−
h
(
u
,
v
)
>
=
<
g
v
Δ
v
,
h
v
Δ
v
>
S
=
∣
A
B
⃗
×
A
C
⃗
∣
=
∣
g
u
h
v
−
g
v
h
u
∣
\vec{AB} = T(u+\Delta u, v) - T(u, v) \\\\ = <g(u+\Delta u, v) - g(u,v), h(u+\Delta u, v) - h(u,v)> \\\\ = <g_u \Delta u, h_u \Delta u> \\\\ \vec{AC} = T(u, v+\Delta v) - T(u, v) \\\\ = <g(u, v+\Delta v) - g(u,v), h(u, v+\Delta v) - h(u,v)> \\\\ = <g_v \Delta v, h_v \Delta v> \\\\ S = |\vec{AB} \times \vec{AC}| = |g_uh_v- g_vh_u|
AB=T(u+Δu,v)−T(u,v)=<g(u+Δu,v)−g(u,v),h(u+Δu,v)−h(u,v)>=<guΔu,huΔu>AC=T(u,v+Δv)−T(u,v)=<g(u,v+Δv)−g(u,v),h(u,v+Δv)−h(u,v)>=<gvΔv,hvΔv>S=∣AB×AC∣=∣guhv−gvhu∣
Definition of Jacobian of the transformation T
Given by
x
=
g
(
u
,
v
)
x = g(u,v)
x=g(u,v) and
y
=
h
(
u
,
v
)
y = h(u,v)
y=h(u,v) is
∂
(
x
,
y
)
∂
(
u
,
v
)
=
g
u
h
v
−
g
v
h
u
\frac{\partial (x,y)}{\partial (u,v)} = g_uh_v- g_vh_u
∂(u,v)∂(x,y)=guhv−gvhu
In the general case, the change of variable in integrals
∫
∫
∫
E
d
V
=
∫
∫
∫
V
∣
∂
(
x
,
y
,
z
)
∂
(
u
,
v
,
w
)
∣
d
u
d
v
d
w
\int\int\int_E dV = \int\int\int_V |\frac{\partial(x,y,z)}{\partial(u,v,w)}| \ dudvdw
∫∫∫EdV=∫∫∫V∣∂(u,v,w)∂(x,y,z)∣ dudvdw
E.x.
We have a region R bounded by (1,0), (2,0), (0,-2), and (0,-1). Evaluate ∫ ∫ R e x + y x − y d A \int\int_R e^{\frac{x+y}{x-y}} dA ∫∫Rex−yx+ydA.
sol:
We first let u = x + y , v = x − y u=x+y, \ v=x-y u=x+y, v=x−y. Verify that this is an one-to-one map.
Compute the Jacobian.
∂
(
x
,
y
)
∂
(
u
,
v
)
=
x
u
y
v
−
x
v
y
u
=
−
1
2
\frac{\partial (x,y)}{\partial (u,v)} = x_uy_v- x_vy_u = -\frac 1 2
∂(u,v)∂(x,y)=xuyv−xvyu=−21
The next step is to figure out how R is transformed under the map T.
(x,y) = (1,0) => (u,v) = (1,1)
(x,y) = (2,0) => (u,v) = (2,2)
(x,y) = (0,-1) => (u,v) = (-1,1)
(x,y) = (0,-2) => (u,v) = (-2,2)
∫
∫
R
e
x
+
y
x
−
y
d
A
=
∫
∫
S
e
u
v
∣
−
1
2
∣
d
u
d
v
=
∫
1
2
∫
−
v
v
1
2
e
u
v
d
u
d
v
=
∫
1
2
[
1
2
v
e
u
v
]
−
v
v
d
v
\int\int_R e^{\frac{x+y}{x-y}} dA = \int\int_S e^{\frac{u}{v}} |-\frac 1 2| dudv \\\\ = \int_{1}^{2}\int_{-v}^{v} \frac 1 2 e^{\frac{u}{v}} dudv \\\\ = \int_{1}^{2} [\frac 1 2 ve^{\frac u v}]_{-v}^{v} dv \\\\
∫∫Rex−yx+ydA=∫∫Sevu∣−21∣dudv=∫12∫−vv21evududv=∫12[21vevu]−vvdv
Reference
- Multivariable_Calculus_8th_Edition (15.1-15.9), James Stewart