S Pital' o H L': WLB 10102 / WLB 10703 /LN/01 Chap 1 Advanced Differentiation
S Pital' o H L': WLB 10102 / WLB 10703 /LN/01 Chap 1 Advanced Differentiation
Objectives:
Step 1. Differentiate both sides of the equation with respect to x . Remember that y is
really a function of x for part of the curve and use the chain rule when
differentiating terms containing y .
dy
Step 2. Solve the differentiated equation algebraically for .
dx
Example 1
x 2 y + 2 y 3 = 3x + 2 y
dy
find .
dx
Solution
d 2
dx
(
x y + 2y3 =
d
dx
)
(3x + 2 y )
d 2
dx
x y +2( )
d 3
dx
d
dx
d
dx
( )
y = 3 (x ) + 2 ( y )
x2
d
( )
( y ) + y d x 2 + 2 3 y 2 d ( y ) = 3(1) + 2 dy
14 dx4424dx 443 1 4 42 dx443 dx
Product rule Extended Power rule
dy dy dy
x2 + 2 xy + 6 y 2 = 3+ 2
dx dx dx
dy
Finally, solve this equation for :
dx
(x 2
+ 6y2 − 2 ) dy
dx
= 3 − 2 xy
dy 3 − 2 xy
∴ = 2 ■
dx x + 6 y 2 − 2
a) 2y4 b) sin 3t
Solution
It is possible to differentiate
a) Let u = 2 y , then, by the function of a function rule:
4 an implicit function by
using the function of a
function rule, which may be
du du dy d dy
= × = (2 y 4 ) × du du dy
dx dy dx dy dx stated as = × .
dx dy dx
dy i.e. u = f(y) .
= 8y3 ■
dx
du du dt d dt
= × = (sin 3t ) ×
dx dt dx dt dx
dt ■
= 3 cos 3t
dx
a) 4 ln 5 y 1 3θ − 2
b) e
5
Solution
du du dy d dy
= × = (4 ln 5 y ) ×
dx dy dx dy dx
4 dy
= ■
y dx
1 3θ − 2
b) Let u = e , then, by the function of a function rule:
5
du du dθ d 1 3θ − 2 dθ
= × = ( e )×
dx dθ dx dθ 5 dx
3 dθ ■
= e 3θ − 2
5 dx
d
Example 4 Determine (2 x 3 y 2 ).
dx
Solution
d d d
(2 x 3 y 2 ) = (2 x 3 ) ( y 2 ) + ( y 2 ) (2 x 3 )
dx dx dx
dy
= (2 x 3 ) 2 y + ( y 2 )(6 x 2 )
dx
dy
= 4x3 y + 6x2 y 2
dx
dy
= 2 x 2 y 2 x + 3 y ■
dx
d 3y
Example 5 Find ( ).
dx 2 x
Solution
d d
(2 x) (3 y ) − (3 y ) (2 x)
d 3y dx dx
( )= 2
dx 2 x (2 x)
dy
(2 x) 3 − (3 y )(2)
= dx
4x 2
dy
6x − 6 y
= dx
4x2
3 dy
= 2 x − y ■
2 x dx
Do it yourself
4. Determine
d dz
a) (3 x 2 y 3 ) c) given z = 2 x 3 ln y
dx dy
d 3u
b) ( )
du 4v
d 2y
5. Find ( )
dx 5 x
∂f ∂z ∂
f x ( x, y ) = = = f ( x, y ) = z x = D x ( f )
∂x ∂x ∂x
and
∂f ∂z ∂
f y ( x, y ) = = = f ( x, y ) = z y = D y ( f )
∂y ∂y ∂y
The values of the partial derivatives of f ( x, y ) at the point (a, b) are denoted by
∂f ∂f
= f x ( a, b) and = f y ( a, b)
∂x ( a ,b ) ∂y ( a ,b )
Example 6 If f ( x, y ) = x 3 y + x 2 y 2 , find
a) fx b) fy
Solution
∂z
a)
∂x π
( ,0 )
3
b) z y at (1, 1).
Solution
∂z
a) = 2 x sin(3 x + y 3 ) + x 2 cos(3 x + y 3 )(3)
∂x
= 2 x sin(3 x + y 3 ) + 3 x 2 cos(3 x + y 3 )
Thus,
∂z
2
π π 2π π2 π2
= 2 sin π + 3 cos π = ( 0) + (−1) = − ■
∂x π
( ,0) 3 3 3 3 3
3
Let f ( x, y, z ) = x 2 + 2 xy 2 + yz 3 ; determine:
a) fx b) fy c) fz
Solution
b) f x ( x, y, z ) = 4 xy + z 3 ■
c) f z ( x, y, z ) = 3 yz 2 ■
Solution
∂z ∂z
2 xz + x 2 + 3 yz 2 =1
∂x ∂x
∂z
Then solve the equation for :
∂x
∂z 1 − 2 xz
= 2
∂x x + 3 yz 2
∂z ∂z
x2 + z 3 + 3 yz 2 =0
∂y ∂y
so that
∂z − z3
= ■
∂y x 2 + 3yz 2
Do it yourself
1. If z = 5 x 4 + 2 x 3 y 2 − 3 y find
∂z ∂z
a) and b) .
∂x ∂y
1 ∂z 1 ∂z
3. If z = sin xy show that = .
y ∂x x ∂y
∂z ∂z
4. In Problems (a) – (f), find and .
∂x ∂y
a) z = 2 xy
b) z = x 3 − 2 xy + y 2
x
c) z =
y
d) z = sin( 4 x + 3 y )
y 1
e) z = x 3 y 2 − 2 +
x y
f) z = cos 3 x sin 4 y
1.2.2 HIGHER-ORDER PARTIAL DERIVATIVES (2ND ORDER & MIXED 2ND ORDER)
Given z = f ( x, y ).
∂2 f ∂ ∂f
= = ( f x ) x = f xx
∂x 2
∂x ∂x
∂2 f ∂ ∂f
= = ( f y ) y = f yy
∂y 2
∂y ∂y
∂2 f ∂ ∂f
= = ( f y ) x = f yx
∂x∂y ∂x ∂y
∂2 f ∂ ∂f
= = ( f x ) y = f xy
∂y∂x ∂y ∂x
Solution
a) First differentiate with respect to y ; then differentiate with respect to x.
∂z
= −2 x + 9 y 2
∂y
∂ z ∂ ∂z ∂
( )
2
= = − 2 x + 9 y 2 = −2 ■
∂x∂y ∂x ∂y ∂x
d) Evaluate the mixed partial found in part (b) at the point (3, 2):
f xy (3, 2) = −2 ■
Solution
We have the partial derivatives
f x = 2 xye y f y = x 2 e y + x 2 ye y
The mixed partial derivatives (which must be the same by the previous theorem) are
f xy = ( f x ) y = 2 xe y + 2 xye y ■ f yx = ( f y ) x = 2 xe y + 2 xye y ■
Solution
∂T x
= −e −t cos
∂t c
and
∂ 2T ∂ 1 x
= − e −t sin
∂x 2
∂x c c
1 x
= − 2 e −t cos
c c
∂T 2 ∂ T
2
Thus, T satisfies the heat equation =c ■
∂t ∂x 2
∂f ∂f
df = dx + dy = f x ( x, y )dx + f y ( x, y )dy
∂x ∂y
where dx and dy are independent variables. Similarly, for a function of three variables
w = f ( x, y, z ) the total differential is
∂f ∂f ∂f
df = dx + dy + dz …………….( ) أ
∂x ∂y ∂z
Solution
∂f ∂f ∂f
a) df = dx + dy + dz = 6 x 2 dx + 20 y 3 dy − 6dz ■
∂x ∂y ∂z
∂f ∂f
b) df = dx + dy
∂x ∂y
−2 2 6y
= 2 x ln(3 y 2 − 2 x) + x 2 2 dx + x dy
3 y − 2x 3y − 2x
2
2x 2 6x 2 y
= 2 x ln(3 y 2 − 2 x) − 2 dx + dy ■
3 y − 2x 3y 2 − 2x
2x
Example 14 If z = f ( x, y ) and z = x 2 y 3 + + 1, determine the total differential, dz.
y
Solution
The total differential is the sum of the partial differentials, i.e.
∂z ∂z
dz = dx + dy
∂x ∂y
∂z 2
= 2 xy 3 + (i.e. y is kept constant)
∂x y
∂z 2x
= 3x 2 y 2 − 2 (i.e. x is kept constant)
∂y y
2 2x
Hence dz = 2 xy 3 + dx + 3 x 2 y 2 − 2 dy ■
y y
Solution
The total differential
∂z ∂z ∂z
dz = du + dv + dw
∂u ∂v ∂w
∂z
= 6u (i.e. v and w are kept constant)
∂u
∂z
= −2 + 8w 3 v (i.e. u and w are kept constant)
∂v
∂z
= 12 w 2 v 2 (i.e. u and v are kept constant)
∂w
Hence dz = 6udu + (8vw 3 − 2)dv + (12v 2 w 2 )dw ■
Do it yourself
f ′( x)
lim =L
x →c g ′( x)
f ( x)
lim =L
x →c g ( x)
The theorem also applies to one-sided limits and to limits at infinity (where x → +∞ and
x → −∞ ).
Solution
Note that this is an indeterminate form because sin x and x both approach 0 as x → 0. This
means that l’Hopital’s rule applies:
d
(sin x)
sin x cos x 1
lim = lim dx = lim = =1 ■
x →0 x x →0 d x →0 1 1
( x)
dx
Solution
For this example, f ( x) = x 7 − 128 and g ( x) = x 3 − 8 , and the form is 0/0.
d 7
( x − 128)
x 7 − 128
lim 3 = lim dx l’Hopital’s rule
x→2 x − 8 x→2 d 3
( x − 8)
dx
7x6
= lim
x→2 3x 2
7x4
= lim Simplify
x→2 3
7(2) 4 112
= = ■ Limit of a quotient
3 3
Solution
You must always remember to check that you have an indeterminate form before applying
l’Hopital’s rule. The limit is
1 − cos x lim (1 − cos x) 0
lim = x →0 = =0 ■
x→0 sec x lim sec x 1
x →0
ATTENTION: If you blindly apply l’Hopital’s rule in Example 19, you obtain the WRONG
answer:
1 − cos x sin x
lim = lim
x →0 sec x x → 0 sec x tan x
cos x 1
= lim = =1
x →0 sec x 1
Solution
This is a 0/0 indeterminate form, and we find that
x − sin x 1 − cos x
lim 3
= lim
x →0 x x →0 3x 2
This is still the indeterminate form 0/0, so l’Hopital’s rule can be applied once again:
1 − cos x − (− sin x) 1 sin x 1 1
lim 2
= lim = lim = (1) = ■
x →0 3x x → 0 6x 6 x → 0 x 6 6
Solution
We could compute this limit by multiplying by 1 x 2 ( ) (1 x ) . Instead, we note that this is of
2
4 2
= lim = ■
x → +∞ 6 3
Solution
The limit has the form 0/0, but direct application of l’Hopital’s rule leads to a real mess.
Instead, we compute the given limit by using the product rule for limits first, followed by two
simple applications of l’Hopital’s rule. Specifically, using the product rule for limits, we have
sin x 4 cos 4 x 1
= lim lim lim
x →0 2 x x →0 cos x
1 x → 0
Solution
This limit has the indeterminate form ∞ ∞ . If you try to apply l’Hopital’s rule, you find
x + sin x 1 + cos x
lim = lim .
x → +∞ x − cos x x → +∞ 1 + sin x
The limit on the right does not exist, because both sin x and cos x oscillate between -1 and 1
f ′( x)
as x → +∞. Recall that l’Hopital’s rule applies only if lim = L or is ± ∞. This does
x →c g ′( x)
not mean that the limit of the original expression does not exist or that we cannot find it; it
simply means that we cannot apply l’Hopital’s rule. To find this limit, factor out an x from the
numerator and denominator and proceed as follows:
sin x sin x
x1 + 1+
x + sin x x = 1+ 0 = 1 ■
= lim
x
lim = lim
x → +∞ x − cos x x → +∞ cos x x → +∞ cos x 1 − 0
x 1 − 1−
x x
x
1
Show that lim 1 + = e.
x → +∞
x
Solution
Note that this limit is indeed of the indeterminate form 1∞ . Let
x
1
L = lim 1 +
x → +∞
x
Take the logarithm of both sides:
1
x
ln L = ln lim 1 +
x → +∞
x
x
1
= lim ln1 + ln x is continuous
x → +∞
x
1
= lim x ln1 + Property of logarithms
x → +∞
x
1
ln1 +
= lim
x
Form 0/0
x → +∞ 1
x
1 1
−
1+1 x x2
= lim l’Hopital’s rule
x → +∞ 1
− 2
x
1
= lim Simplify
x → +∞ 1
1+
x
1
=
1+ 0
=1
Thus, ln L = 1 and L = e1 = e ■
Solution
The limit has the form 0 ⋅ ∞ , because
π
lim − x − = 0 and lim tan x = +∞
x→ π 2 2 x→π 2−
1
Write tan x = to obtain
cot x
π
x−
π 2
lim − x − tan x = lim − Form 0/0
x→ π 2 2 x → π 2 cot x
1
= lim − l’Hopital’s rule
x → π 2 − csc x
2
= lim − (− sin 2 x) = −1 ■
x→π 2
Solution
This is a 0 0 indeterminate form. We begin by using properties of logarithms.
L = lim+ x sin x
x →0
ln L = ln lim+ x sin x
x →0
ln x
= lim+ This is ∞ ∞ form
x →0 csc x
1x
= lim+ l’Hopital’s rule
x → 0 − csc x cot x
− sin 2 x
= lim+
x → 0 x cos x
sin x − sin x
= lim+
x →0 x cos x
= (1)(0) = 0
Thus, L = e0 = 1 ■
Solution
This is a limit of the indeterminate form ∞ 0 .
If L = lim x1 x , then
x → +∞
= lim ln x1 x limit
x → +∞
1
= lim
ln x
x → +∞ x
ln x ∞
= lim This is
x → +∞ x ∞
1
= lim x l’Hopital’s rule
x → +∞ 1
=0
Thus, we have ln L = 0; therefore, L = e 0 = 1 ■
Solution
1 1
As it stands, this has the form ∞ − ∞ , because → +∞ and → +∞ as x → 0 from
x sin x
the right. However, using a little algebra, we find
1 1 sin x − x
lim+ − = lim+
x→0 x sin x x →0 x sin x
This limit is now of the form 0/0, so the hypotheses of l’Hopital’s rule are satisfied.
Thus,
sin x − x cos x − 1
lim+ = lim+ This is 0/0 form
x→0 x sin x x →0 sin x + x cos x
− sin x 0
= lim+ = =0 ■
x →0 cos x + x ( − sin x ) + cos x 2
ATTENTION: Not all limits that appear indeterminate actually are indeterminate. For
example,
lim+ (sin x)1 x = 0 0 ∞ form
x→0
tan x 0
lim+ =0 form
x→0 ln x ∞
Other such “false indeterminate forms” include + ∞ + (+∞), ∞ 0 , and ∞.∞, which are all
actually infinite.
Do it yourself
1 − cos x sin x
a) lim = lim =0
x→π x x→π 1
sin x cos x
b) lim = lim =0
x→ π 2 x x→ π 2 1
x10 − 1 1
b) lim h) lim − + tan x
x →1 x − 1 x→( π 2 ) π − 2 x
x 2 + 3x − 4
3x
1
c) lim 2 i) lim 1 +
x →1 x − 7 x + 6 x → +∞
2x
1 − cos 2 x ln(ln x)
d) lim j) lim
x→0 sin 3 x x → +∞ x
e) lim
x − sin x
x →0 tan x − x
k) (
lim+ e x + x
x→0
)
1x
sin x − x 1
f) lim l) lim+ 2 − ln x
x→0 x
x →0 x2
Solution
We can write R = k ( pl − p 2 ). Setting dR dp = 0 and solving for p gives
dR l
= k (l − 2 p ) = 0, ∴ p = .
dp 2
Now, d R dp = −2k . Since k > 0, the second derivative is always negative. Hence,
2 2
Solution
To locate the extrema, we solve C ′(t ) = 0.
d k
C ′(t ) = (e − at − e −bt )
dt b − a
=
k
b−a
[
(− a )e − at − (−b)e −bt =
k
]
b−a
(be −bt − ae − at )
k
lim C (t ) = lim (e − at − e −bt )
t → +∞ t → +∞ b − a
k 1 1
= tlim − lim bt
b−a → +∞ e at t → +∞ e
=
k
(0 − 0)
b−a
=0 ■
Solution
P( x) = Ax s e − sx r
s
P′( x) = Ax s − e − sx r + e − sx r .sAx s −1 = 0
r
sA s − sx r
xe = sAx s −1e − sx r
r
xs
= x s −1
r
xs
=r
x s −1
x s − s +1 = r
∴x = r ■
Example 32 Beehives are formed by packing together cells that may be modeled as
regular hexagonal prisms open at one end. It can be shown that a cell with hexagonal
side of length s and prism height h has surface area
S (θ) = 6 sh + 1.5s 2 (− cot θ + 3 csc θ)
π
for 0 < θ < . What is the angle θ (to the nearest degree) that minimizes the surface
2
area of the cell (assuming that s and h are fixed)?
Solution
S (θ) = 6 sh + 1.5s 2 (− cot θ + 3 csc θ)
S ′(θ) = 1.5s 2 (− csc 2 θ − 3 csc θ cot θ) = 0
csc 2 θ = 3 csc θ cot θ
csc θ = 3 cot θ
1 cos θ
= 3
sin θ sin θ
1
cos θ =
3
∴ θ = 54.7 o ≈ 55 o ■
df ∂f dx ∂f dy ∂f dz
= + + ……()ب
dt ∂x dt ∂y dt ∂z dt
Example 33 The height of a right circular cone is increasing at 3 mm/s and its
radius is decreasing at 2 mm/s. Determine, correct to 3 significant figures, the rate at
which the volume is changing (in cm3/s) when the height is 3.2 cm and the radius is
1.5 cm.
Solution
1 2
Volume of a right circular cone, V = πr h
3
Using equation ()ب, the rate of change of volume,
dV ∂V dr ∂V dh
= +
dt ∂r dt ∂h dt
∂V 2 ∂V 1 2
= πrh and = πr
∂r 3 ∂h 3
dh
Since the height is increasing at 3 mm/s, i.e. 0.3 cm/s, then = + 0 .3
dt
dr
and since the radius is decreasing at 2 mm/s, i.e 0.2 cm/s, then = − 0 .2 .
dt
Hence
dV 2 1
= πrh (− 0.2 ) + πr 2 (+ 0.3)
dt 3 3
− 0 .4
= πrh + 0.1πr 2
3
1
Example 34 The area A of a triangle is given by A = ac sin B, where B is the
2
angle between sides a and c. If a is increasing at 0.4 units/s, c is decreasing at 0.8
units/s and B is increasing at 0.2 units/s, find the rate of change of the area of the
triangle, correct to 3 significant figures, when a is 3 units, c is 4 units and B is π 6
radians.
Solution
Using equation()ب, the rate of change of area,
dA ∂A da ∂A dc ∂A dB
= + +
dt ∂a dt ∂c dt ∂B dt
1 ∂A 1
Since A = ac sin B, = c sin B ,
2 ∂a 2
∂A 1 ∂A 1
= a sin B and = ac cos B
∂c 2 ∂B 2
da dc dB
= 0.4 units/s, = −0.8 units/s and = 0.2 units/s
dt dt dt
Hence
dA 1 1 1
= c sin B (0.4) + a sin B (−0.8) + ac cos B (0.2)
dt 2 2 2
π
When a = 3, c = 4 and B = then:
6
dA 1 π 1 π 1 π
= (4) sin (0.4) + (3) sin (−0.8) + (3)(4) cos (0.2)
dt 2 6 2 6 2 6
= 0.839 units / s ■ (correct to 3 significant figures)
2
It is often useful to find an approximate value for the change (or error) of a quantity caused by
small changes (or errors) in the variables associated with the quantity. If w = f ( x, y, z ) and
δx, δy, δz denote small changes in x, y, z respectively, then the corresponding
approximate change δf in f is obtained from equation ( ) أby replacing the
differentials by the small changes.
Thus
∂f ∂f ∂f
δf ≈ δx + δy + δz …………….( )
∂x ∂y ∂z
Solution
Using equation ( ), the approximate error in k,
∂k ∂k
δk ≈ δp + δV
∂p ∂V
Let p, V and k refer to the initial values.
∂k ∂k
Since k = pV 1.4 then = V 1.4 and = 1.4V 0.4 .
∂p ∂V
4
Since the pressure is increased by 4%, the change in pressure δp = × p = 0.04 p.
100
Solution
∂G ∂G ∂G
Using δG ≈ δR + δθ + δL .
∂R ∂θ ∂L
( R 4 θ) ∂G 4 R 3 θ ∂G R 4 ∂G − R 4 θ
Since G = , = , = and = .
L ∂R L ∂θ L ∂L L2
2
Since R is increased by 2%, δR = R = 0.02 R . Similarly,
100
δθ = −0.05θ and δL = 0.04 L.
Hence
4 R 3θ R4 − R 4θ
δG ≈ (0.02 R ) + (− 0.05θ ) + 2 (0.04 L )
L L L
R 4θ 1
≈ −0.01 , i.e. δG ≈ − G ■
L 100
Hence the approximate percentage error in G a 1% decrease.
Definition
f ( xo + hu1 , yo + hu2 ) − f ( xo , yo )
Du f ( xo , yo ) = lim
h→0 h
Solution
1 3
Du f (1, 2) = f x (1, 2) + f y (1, 2) −
2 2
1 3
= −4(1) + 3(2) 2 − = −2 − 6 3 ≈ −12.4 ■
2 2
The Gradient
∇f ( x, y ) = f x ( x, y )i + f y ( x, y ) j
∇f o = f x ( xo , yo )i + f y ( xo , yo ) j
Solution
∂ 2 ∂
f x ( x, y ) = ( x y + y 3 ) = 2 xy and f y ( x, y ) = ( x 2 y + y 3 ) = x 2 + 3 y 2
∂x ∂y
Then ∇f ( x, y ) = 2 xyi + ( x 2 + 3 y 2 ) j ■
Solution
2x 2
f x ( x, y ) = , so f x (1, - 3) = −
x + y3
2
26
3y2 27
f y ( x, y ) = , so f y (1, - 3) = −
x +y
2 3
26
2 27
∇f o = ∇f (1,−3) = − i− j
26 26
v 2i − 3 j 1
u= = = (2i − 3 j )
v 2 + (−3)
2 2
13
Thus,
2 2 27 3
Du ( x, y ) = ∇f .u = − + − −
26 13 26 13
77 13
= ■
338
Solution
The heat flow Hˆ = −k∇T where k = the thermal conductivity (a positive constant)
From T ( x, y ) = x 2 + y 2 , we have ∇T = 2 xi + 2 y j .
Because the thermal conductivity k is positive, we can say that heat flows from Po in
the direction of the unit vector u given by
− (6i + 8 j ) 3 4
u= =− i− j ■
(−6) 2 + (−8) 2 5 5