Lecture 4
Author: Kemal Ahmed Instructor: Dr. French Course: Math 2ZZ3 Date: 2013-07-08Math objects made using MathType; figures taken from Advanced Engineering Math 4th Edition (Zill); graphs made with Winplot.
Announcements
• Test will be July 17th for sure
• No formula sheet will be given on the test
• As much as French hates our textbook, still use it. Also, use your first year Stewart Calculus textbook.
9.3 Review
e.g. 1)Find aN,aT.
( )
t = cosh( )
t , sinh( )
tr
Recall:
( )
( )
( )
( )
( )
( )
cosh
2
sinh
2 d
cosh sinh d
d
sinh cosh
d
t t
t t
e e t
e e t
t t
t
t t
t
−
−
+ =
− =
=
=
Back to the question:
( )
( )
( )
( )
( )
( )
( )
( )
sinh , cosh cosh , sinh
t t t t
t t t t
′ = =
′′ = =
r v
r a
( )
( )
2 sinh t cosh t ′ ′′
⋅ = ⋅ =
v a r r
( )
( )
2 2
0, 0, sinh t cosh t 0, 0, 1
× = − = −
v a
1 ⇒ × =v a
( )
( )
( )
( )
2 2
2 sinh cosh sinh cosh T
T a
t t
a
t t
⋅ =
′ ′′⋅ =
′ =
+
v a v
r r r
( )
( )
2 2
1
sinh cosh N
a
t t
× =
′ ′′× =
′
=
+
v a v r r
r
e.g. 2)
Find the equation of the osculating plane,r
( )
t = 2 cos( )
t , 2 sin( )
t , 3t at 4 t=π .We need a point and a normal vector. The point is given by:
( )
4( )
4( ) ( )
4 43 4
2 cos , 2 sin , 3 2, 2,
π π π π
π =
=
r
T, N span the osculating plane.∴ ×T Nis⊥to osculating plane
B will suffice as a normal vector.
( )
( )
( )
( )
2( )
2( )
2 sin , 2 cos , 3 4 sin 4 cos 9 13
t t t
t t t
′ = −
′ = + +
=
r
r
( )
( )
( )
1( )
( )
2 sin , 2 cos , 3 13
t
t t t
t ′
= = −
′
r T
r
( )
1( )
( )
2 cos , 2 sin , 0 13
t t t
′ = − −
T
( )
4 23 13
t
′ = =
T
( )
t = −2 cos( )
t , 2 sin−( )
t , 0N
( )
( )
( )
( )
4 4 4
1 1
4 2 2
2 cos , 2 sin , 0 , , 0
π π π
π
= − −
= − −
N
N
3 3 2
, ,
4 4 4 26 26 13
π π π −
= × =
B T N
Take an arbitrary point on the osculating plane,
(
x y z, ,)
. 32, 2,
4
x y z π
⇒ − − − is in the plane 3
2, 2,
4
3 3 2 3
, , 2, 2, 0
4 26 26 13
x y z
x y z
π
π
⊥ − − −
−
⋅ − − − =
B
3
3 3 2 2
2 x− y+ z= π
9.4 – Partial Derivatives
( )
,z= f x y
2
:
f →
The graph of f consists of points in3like
(
x y f x y, ,( )
,)
and describes a surface in3.Level Curves
Imagine cutting graph of the f by successive horizontal planes.
Pick some k in the range of f, such thatz=k.
Every point on the intersection of surface and the plane looks like
(
x y k0, 0,)
, where(
0, 0)
f x y =k.
This is the basis of contour mapping.
Level Curves of f are the curves with equations f x y
( )
, =k, where k ϵ Range of f.e.g. 3)
Sketch level curves of f x y
( )
, = 9−x2−y2 To find the range, find the domain.2 2 2 2
9−x −x ≥ ⇔ ≥0 9 x +y
Thus, the domain is a disk with equationx2+y2 =9
( )
( )
2 2
2 2
0 , 3
9 , 0
x y f x y
x y f x y
+ = ⇒ =
+ = ⇒ =
Range:
[ ]
0, 3 Try k = 0( )
2 2
2 2
, 0
9 9 f x y
x y x y
=
= − −
+ =
k = 1
( )
2 2
2 2
, 1
9 8 f x y
x y x y
=
= − −
+ =
k = 2
( )
2 2
2 2
, 2
9 5 f x y
x y x y
=
= − −
+ =
k = 3
( )
( ) ( )
2 2
2 2
, 3
9 0
, 0, 0 f x y
x y x y
x y =
= − −
+ =
⇔ =
Note: example is from Stewart’s 14.1, example 11
(
)
3
, , :
w f x y z F
= →
Graph is a set of points in 4
,
(
)
independent dependent
, , , , , x y z F x y z
Graph is 4D
(
, ,)
,F x y z =k k∈Range of F This set forms a surface in3.
Setting w = k gives the level surface of F.
e.g. 4)
(
)
2 2 2, , 3 6
F x y z =x + y + z Draw some level surfaces. Any k ≥ 0 makes sense. k = 0
(
)
(
) (
)
2 2 2
, , 0
3 6
, , 0, 0, 0
F x y z
x y z
x y z
=
= + +
⇔ =
k = 1
(
)
2 2 2, , 1 3 6
F x y z = =x + y + z ←ellipsoid k = 2
(
)
2 2 2, , 2 3 6
F x y z = =x + y + z ←ellipsoid Partial Derivatives
Recall:z= f x y
( )
,(
)
( )
0, ,
lim x
f x x y f x y z
x ∆ → x
+ ∆ −
∂ =
∂ ∆
= partial derivative with respect to x
(
)
( )
0
, , ,
lim y
f x y y f x y z
y ∆ → y
∆ −
∂ =
∂ ∆
= partial derivative with respect to y
Find the derivative with respect to one variable and hold the others constant.
e.g. 5)
( )
3 2 3 2
,
6 5
f x y z
z x x y y
=
= − + +
2 3
2 2
3 12 0
0 18 10
z
x xy
x z
x y y
y
∂ = − + +
∂
∂ = + +
∂
e.g. 6)
( )
( )
2tan1 2
,
x y
f x y z z e
−
= =
( )
( )
( )
2 1 2
2 1 2
tan 1 2
tan 2
4
2 tan
2 1
x y
x y
z
e x y
x
z y
e x
y y
−
−
−
∂ =
∂
∂
=
∂ +
x x z f
f z x x
∂ = ∂ = =
∂ ∂
( )
,f x y =z ,
f f x y ∂ ∂
∂ ∂ are also functions of 2 variables
2 →
So you may differentiate fx,fyagain with respect to either x or y.
e.g. 7)
3 2 3 2
6 5
x x y y
− + +
2 3
2
2
2 3
3
3 12
3 12
6 12 z
x xy x
z z
x x x
x xy x
x y
∂ = − + ∂
∂ = ∂ ∂ ∂ ∂ ∂
∂
= − +
∂ = − + OR
2
2 3
2
3 12
36
z x
y x y x
x xy
y xy
∂ = ∂ ∂
∂ ∂ ∂ ∂
∂
= − +
∂ =
(2nd mixed partial) Notation
nd st 3
2
1 2
z y x ∂ ∂ ∂
rd nd st 3
3 2 1
z x y z
∂ ∂ ∂ ∂
st 1 2
x y z
f y x ∂
= ∂ ∂
The order is because it is intuitive to read fxyand differentiate it in order that it appears and the order for the other notation is read right-left because you split up your derivatives from right-left, like so:
2
z z
y x x y
∂ ∂ ∂
=
∂ ∂ ∂ ∂ .
e.g. 7 continued
Continued from example 7.
3 2 3 2
6 5
x + x y + y =z
2
2
2 2
2
36
18 10
36 z
xy y x
z x y
x y y x
xy
∂ =
∂ ∂
∂ ∂
=
∂ ∂
∂
= +
∂ =
f has continuous 2nd order partials => fxy = fyx
For 3/+ Variables
Do everything the same, but with more independent variables.
(
, ,)
w=F x y z 1st partials
, , F F F
x y z
∂ ∂ ∂
∂ ∂ ∂
Chain Rule
The best way to learn this is to practice like crazy.
Version 1: 2D
Ifz= f x y
( )
, ,u=g x y( )
, ,v=h x y( )
, and all have continuous 1st order partials, then :z z u z v
x u x v x
∂ = ∂ ∂⋅ +∂ ∂⋅
∂ ∂ ∂ ∂ ∂
Note: x:=ymeans x is defined by y. z z u z v
y u y v y
∂ ∂ ∂ ∂ ∂
= ⋅ + ⋅
∂ ∂ ∂ ∂ ∂
(
1, 2, 3,...)
, iz= f u u u u uare functions ofX X1, 2,X3,...,Xk
3
1 2
1 2 3
... n
j j j j n j
u u
u u
z z z z z
x u x u x u x u x
∂ ∂
∂ ∂
∂ = ∂ ⋅ + ∂ ⋅ + ∂ ⋅ + + ∂ ⋅
∂ ∂ ∂ ∂ ∂ ∂ ∂ ∂ ∂
e.g. 8)
2 2
1
2 2
tan u r s
w uv
v r s − = −
=
=
Compute:
(
)
( )
1 1
2 2 2
1 1 1 1
2 2
1 2 1 2
w w u w v
s u s v s
u v s v u r s
uv uv
− −
∂ =∂ ⋅∂ +∂ ⋅∂
∂ ∂ ∂ ∂ ∂
= ⋅ ⋅ − + ⋅ ⋅
+ +
9.5 – Directional Derivatives
d& d x x ∂
∂ are not functions, but rather, differential operators think of them as “verbs”
2D
, x y ∂ ∂ ∇ =
∂ ∂
3D , , x y z ∂ ∂ ∂ ∇ =
∂ ∂ ∂
3
:
F →
3 3
, ,
:
F F F
F
x y z
F
∂ ∂ ∂
∇ =
∂ ∂ ∂
∇ →
∇= “nabla” f
∇ = gradient of f = grad(f) e.g. 9)
(
)
( )
(
)
( )
( )
( )
2( )
, , cos
, , , ,
cos , cos sin , sin
x y z F x y z xy yz F x y z F F F
y yz x yz xy yz z xy yz =
∇ =
= − −
The gradient is a vector quantity. Directional derivatives
Since the unit vectoruˆ = cos , sinθ θ ,z= f x y
( )
, , the directional derivative of f in direction ˆuis:
( )
(
)
( )
0
cos , sin ,
, lim
u
h
f x h y h f x y D f x y
h
θ θ
→
+ + −
=
Note: θ = 0, definition yields fx
2
π
θ = , definition yields fy Theorem:
( )
( )
( )
ˆ
, , cos , sin ˆ
, ,
u
z f x y u D f x y f x y
θ θ
= =
= ∇ ⋅u
ˆ
u
must be of unit length!
If you want directional derivative in direction v, where v is not unit length, normalize v before applying the theoremuˆ = v
v
Proof
( )
(
cos , sin)
g t = f x t+ θ y t+ θ , ,
x y θfixed, g is function of 1-variable Computeg′
( )
0 by 2 methods:Method 1: Limit Definition
( )
(
) ( )
(
)
( )
0
0
0 0
0 lim
cos , sin ,
lim h
h
u
g h g
g
h
f x h y h f x y
h D f
θ θ
→
→
+ −
′ =
+ + −
=
=
( )
1(
)
[
]
2(
)
[
]
d d
cos , sin cos cos , sin sin
d d
g t f x t y t x t f x t y t y t
t t
θ θ θ θ θ θ
′ = + + ⋅ + + + + ⋅ +
1, 2
f f are partial derivatives of f x t
(
+ cos ,θ y t+ sinθ)
with respect to cos , sin ,u v
x t+ θ y+t θ respectively
[
]
[
]
( )
d d
cos sin
d d
cos sin
u v
u v
f x t f y t
t t
f f
g t
θ θ
θ θ
= + + +
= +
′ =
Set t = 0 => u=x v, = y
( )
( )
( )
0 cos sin
, cos , sin ,
,
x y
x y
u
g f f
f f f x y u D f x y
θ θ
θ θ
′ = +
= ⋅
= ∇ ⋅
=
Method 2: 3/+ dimensions
(
)
(
)
0(
)
(
)
, ,
cos , sin , cos , ,
, , lim u
h
w F x y z
F x h y h z h F x y z
D F x y z
h
α β γ
→
=
+ + + −
=
α,β,γ, are angles between u and coordinate axes Theorem: D F x y zu
(
, ,)
= ∇F x y z(
, ,)
⋅uˆˆ
umust be unit length. If not, normalize
e.g. 10)
( )
1(
)
, tan y @ 2, 2 f x y
x −
= −
and find directional derivative in direction 1, 3−
( )
( )
( )
( ) (
)
(
)
2 2 2
1 1 1
, , ,
1 1
, 2, 2
1 1 1 1 1 1
2, 2 , ,
2 2 2 2 4 4
x y
y y
x x
y f x y f f
x x
x y
f
∇ = = ⋅ − ⋅
+ +
= −
∇ − = ⋅ =
(
)
1 12, 2 , 1, 3
4 4 1 3 4 4
1 2 u
D f − = ⋅ −
= −
− =
Wrong because she didn’t normalize! So let’s normalize first. 1, 3 : 1, 3 1 9 10
1
ˆ 1, 3 10
− − = + =
− =u
(
)
1 1 1 32, 2 , ,
4 4 10 10
1 3
4 10 4 10 1
2 10 u
D f − = ⋅ −
= −
− =
What is the directional derivative for a unit vector that is tangent to a level curve? 0 because it is a level curve, so the gradient is not changing.
( )
,f x y =z, ˆua unit vector
( )
,( )
, ˆ( )
, ˆ cosu
D f x y =Df x y ⋅ = ∇u f x y u φ
ϕ is the angle between ∇f x y
( )
, and ˆuis 0≤ ≤φ π( )
( )
( )
1 cos 1
, u , ,
f x y D f x y f x y φ
⇒ − ≤ ≤
⇒ ∇ ≤ ≤ ∇
( )
,( )
,u
D f x y = − ∇f x y when ϕ = π; i.e. when ˆuand∇f x y
( )
, are opposite( )
,( )
,u
D f x y = ∇f x y when ϕ = 0, i.e. when ˆuand∇f x y
( )
, point in the same directionMax of directional derivative is equal to ∇f and occurs when ˆuis in the same direction as∇f Min of directional derivative is equal to− ∇f , and it occurs when ˆuis in the opposite direction of∇f .
Also, holds in higher dimensions e.g. 11)
Find a vector that gives the direction of maximal increase at the given point, and find the rate.
( )
2( )
, sin @ 0,
4 x
f x y =e y π
( )
( )
( )
( )
( )
( )
2 2
0 0
4 4 4
, ,
2 sin , cos
0, 2 sin , cos
1 2,
2 x y
x x
f x y f f
e y e y
f π e π e π
∇ =
=
∇ =
=
Move in that direction.
Maximal rate of increase =
( )
1 54 2 2
0, 2
f π
∇ = + =