Math 2011-Introduction to Multivariable Calculus (Edited by Dr.

Hon-Ming HO)
Practice Exercises 8: Partial Differentiation

Exercise 1:

𝜕𝑓 𝜕𝑓
Find partial derivatives and of the following functions:
𝜕𝑥 𝜕𝑦

𝑦
1) 𝑧 = 𝑓(𝑥, 𝑦) = tan−1 ( )
𝑥
2) 𝑧 = 𝑓(𝑥, 𝑦) = 𝑒 −𝑥 sin(𝑥 + 𝑦)
3) 𝑧 = 𝑓(𝑥, 𝑦) = 𝑐𝑜𝑠 2 (3𝑥 − 𝑦 2 )
4) 𝑧 = 𝑓(𝑥, 𝑦) = 𝑙𝑜𝑔𝑦 𝑥

Solution to Exercise 1:

Things to do when finding partial derivatives:

1) Identify which variables are independent variables. This 1st step is especially important. How do we know which variables are
independent variables? In this example, we are given that 𝑧 = 𝑓(𝑥, 𝑦) . Thus, symbols 𝑥 and 𝑦 are supposed to be two
independent variables of the given function. The symbol 𝑧 represents the dependent variable which is controlled by the two
independent variables.
2) Look at the symbols of the desired partial derivatives. Identify which variable is the variable of differentiation and
which variable(s) is (are) treated as fixed constant(s) when performing partial differentiation. In this example, we are required
to find two partial derivatives:
𝜕𝑓
→ 𝑥 is the variable of differentiation and all other vartiable(s) is (are) treated as constant(s).
𝜕𝑥
𝜕𝑓
→ 𝑦 is the variable of differentiation and all other vartiable(s) is (are) treated as constant(s).
𝜕𝑦
𝜕𝑓
If the given functions have three independent variables, say 𝑝 , 𝑞 , 𝑟 , then when finding , we need to treat all other
𝜕𝑝
independent variables 𝑞 and 𝑟 as two fixed constants. And so on.
3) Performing partial differentiation is a kind of pattern recognition, recognizing which differentiation rules are to be applied.
Recognize the pattern of the working formula of the given function 𝑧 = 𝑓(𝑥, 𝑦) = ⋯ ⋯ ⋯ ⋯ ⋯ .

(Constant Multiple Rule)

(General power rule)
𝑘♦ ∆
ῧ
(General logarithmic rule)
(General exponential rule)
?
𝒍𝒏 | ῧ | ? ? 𝒆ῧ
▭1
?
(÷ Rule) ▭1 + ▭2 (+ Rule) ?
▭2 ?
? 𝒔𝒊𝒏−𝟏 ῧ 𝐬𝐢𝐧 ῧ (General sine rule)

? ? ?
𝒄𝒐𝒔−𝟏 ῧ ? ? 𝐜𝐨𝐬 ῧ (General cosine rule)
(▭1 ) ∙ (▭2 ) ?
𝐭𝐚𝐧−𝟏 ῧ 𝐭𝐚𝐧 ῧ
(× Rule)

▭1 and ▭2 are expressions in terms of 𝑥 and 𝑦 . ῧ , ῧ , ῧ , ῧ , ῧ , ῧ , ῧ , ῧ , ῧ are expressions in terms of 𝑥 and 𝑦 .

𝑦 𝒚
Part 1): We observe that 𝑧 = 𝑓(𝑥, 𝑦) = tan−1 ( ) = 𝐭𝐚𝐧−𝟏 ῧ where ῧ =
𝑥 𝒙
. We apply the general arctangent rule for
differentiation as follows:
𝜕𝑓 𝜕 1 𝜕 1 𝜕 𝒚 1 −𝑦 1 −𝑦 −𝑦
= [𝐭𝐚𝐧−𝟏 ῧ ] = ( 2 ) ∙ [ῧ]=( 2 )∙ [ ]=( 2 )∙ 2 =( 2) ∙ = 2 .
𝜕𝒙 𝜕𝒙 𝜕𝒙 𝜕𝒙 𝒙 𝑥 𝒚 𝑥 2 𝑥 + 𝑦2
1+ ῧ 1+ ῧ 1+ ῧ 1+( )
𝒙

Practice Exercises 8 (edited by Dr. Hon-Ming HO)-page 1

 𝜕𝑓 𝜕 1 𝜕 1 𝜕 𝒚 1 1 1 1 𝑥
= [𝐭𝐚𝐧−𝟏 ῧ ] = ( 2) ∙ [ῧ]=( 2) ∙ [ ]=( 2) ∙ =( 2) ∙ = 2 .
𝜕𝒚 𝜕𝒚 𝜕𝒚 𝜕𝒚 𝒙 𝑥 𝒚 𝑥 𝑥 + 𝑦2
1+ ῧ 1+ ῧ 1+ ῧ 1+( )
𝒙

Part 2): We observe that 𝑧 = 𝑓(𝑥, 𝑦) = 𝑒 −𝑥 sin(𝑥 + 𝑦) = (▭1 ) ∙ (▭2 ) is a product of two expressions in terms of 𝑥 and
𝑦 . We apply the product rule for differentiation as follows:
𝜕𝑓 𝜕 𝜕 𝜕 𝜕 𝜕
= [(▭1 ) ∙ (▭2 )] = (▭1 ) ∙ [▭2 ] + (▭2 ) ∙ [▭1 ] = (𝑒 −𝑥 ) [sin(𝑥 + 𝑦)] + (sin(𝑥 + 𝑦)) [𝑒 −𝑥 ] .
𝜕𝒙 𝜕𝒙 𝜕𝒙 𝜕𝒙 𝜕𝒙 𝜕𝒙

We observe that sin(𝑥 + 𝑦) = 𝐬𝐢𝐧 ῧ and 𝑒 −𝑥 = 𝒆 ῧ . We apply the general sine and exponential rules for differentiation
as follows:
𝜕 𝜕 𝜕 𝜕
[sin(𝑥 + 𝑦)] = [𝐬𝐢𝐧 ῧ ] = (cos ῧ ) ∙ [ ῧ ] = (cos ῧ ) ∙ [𝑥 + 𝑦] = cos(𝑥 + 𝑦) .
𝜕𝒙 𝜕𝒙 𝜕𝒙 𝜕𝒙
𝜕 −𝑥 𝜕 𝜕 𝜕
[𝑒 ] = [𝒆 ῧ ] = 𝑒 ῧ [ῧ]=𝑒ῧ [−𝑥] = −𝑒 −𝑥 .
𝜕𝒙 𝜕𝒙 𝜕𝒙 𝜕𝒙
Thus we have
𝜕𝑓 𝜕 𝜕
= (𝑒 −𝑥 ) [sin(𝑥 + 𝑦)] + (sin(𝑥 + 𝑦)) [𝑒 −𝑥 ] = 𝑒 −𝑥 cos(𝑥 + 𝑦) + (sin(𝑥 + 𝑦))(−𝑒 −𝑥 ) .
𝜕𝒙 𝜕𝒙 𝜕𝒙
𝜕𝑓 −𝑥
𝜕 𝜕 −𝑥
= (𝑒 ) [sin(𝑥 + 𝑦)] + (sin(𝑥 + 𝑦)) [𝑒 ] = 𝑒 −𝑥 cos(𝑥 + 𝑦) .
𝜕𝒚 𝜕𝒚 𝜕𝒚

∆
Part 3): We observe that 𝑧 = 𝑓(𝑥, 𝑦) = 𝑐𝑜𝑠 2 (3𝑥 − 𝑦 2 ) = ῧ where ῧ = cos(3𝑥 − 𝑦 2 ) and ∆= 𝟐 . We apply the
general power rule for differentiation as follows:
𝜕𝑓 𝜕 𝜕 ∆ ∆−𝟏 𝜕 𝜕
= [𝑐𝑜𝑠 2 (3𝑥 − 𝑦 2 )] = [ῧ ]=∆ ῧ ∙ [ ῧ ] = 2 cos(3𝑥 − 𝑦 2 ) ∙ [cos(3𝑥 − 𝑦 2 )] .
𝜕𝒙 𝜕𝒙 𝜕𝒙 𝜕𝒙 𝜕𝒙

We also observe that cos(3𝑥 − 𝑦 2 ) = 𝐜𝐨𝐬 ῧ where ῧ = 3𝑥 − 𝑦 2 . Then we apply the general cosine rule for
differentiation as follows:
𝜕 𝜕 𝜕
[cos(3𝑥 − 𝑦 2 )] = [𝐜𝐨𝐬 ῧ ] = (− sin ῧ ) ∙ [ ῧ ] = −3 sin(3𝑥 − 𝑦 2 ) .
𝜕𝒙 𝜕𝒙 𝜕𝒙
Thus we have
𝜕𝑓 𝜕
= [𝑐𝑜𝑠 2 (3𝑥 − 𝑦 2 )] = −6 cos(3𝑥 − 𝑦 2 ) sin(3𝑥 − 𝑦 2 ) .
𝜕𝒙 𝜕𝒙

𝜕𝑓 𝜕
= [𝑐𝑜𝑠 2 (3𝑥 − 𝑦 2 )] = 4𝑦 cos(3𝑥 − 𝑦 2 ) sin(3𝑥 − 𝑦 2 ) .
𝜕𝒚 𝜕𝒚

𝜕𝑓 1 𝜕𝑓 − ln 𝑥
Part 4): We have 𝜕𝒙
=
𝑥 ln 𝑦
and
𝜕𝒚
=
𝑦(ln 𝑦)2
.

Exercise 2:
𝒛

The plane 𝑥 = 1 intersects the paraboloid 𝑧 = 𝑓(𝑥, 𝑦) = 𝑥 2 + 𝑦 2 in a 𝒛 = 𝒙𝟐 + 𝒚𝟐

parabola. See the figure on the right. Find the slope of the tangent line to
the intersecting parabola at point (1 ,2 ,5) .

Solution to Exercise 2: Plane 𝒙 = 𝟏

An initial preparation:

1) Look at the equation 𝑧 = 𝑓(𝑥, 𝑦) = 𝑥 2 + 𝑦 2 . Variables 𝑥 and 𝑦

are supposed to be the two independent variables of the function
𝟏
𝑓.
2) About the point (1 ,2 ,5) : Since it is a point lying on the paraboloid, 𝒙 𝒚

the three coordinates must satisfy the equation of the paraboloid

𝑧 = 𝑓(𝑥, 𝑦) = 𝑥 2 + 𝑦 2 . Check 5 = 𝑓(1,2) = 12 + 22 .

Practice Exercises 8 (edited by Dr. Hon-Ming HO)-page 2

 3) The brown curve in the figure above is the intersecting curve of the 𝒛
cutting plane 𝑥 = 1 and the paraboloid whose equations are given
by
𝑧 = 12 + 𝑦 2 𝑎𝑛𝑑 𝑥 = 1 . Plane 𝒙 = 𝟏

Method 1: We do calculus on the 2-dimensional cutting plane 𝑥 = 1 .

On the cutting plane 𝑥 = 1 , what quantities are allowed to vary are the (1 ,2 ,5)
variables 𝑦 and 𝑧 . On the cutting plane 𝑥 = 1 , the variable 𝑦 controls
the variable 𝑧 . So on the cutting plane 𝑥 = 1 , we deal with a function
of a single variable 𝑦 , i.e. 𝑧 = 𝑓(1 , 𝑦) . Thus the slope of the purple
tangent line at the point (1 , 𝟐 ,5) is given by
𝒚

𝑑 𝑑 2
[𝑓(1 , 𝑦)]| = [1 + 𝑦 2 ]| = 2𝑦 |𝑦=𝟐 = 4 . The purple straight line is the tangent line to the intersecting
𝑑𝑦 𝑦=𝟐
𝑑𝑦 𝑦=𝟐 brown curve at point (1 ,2 ,5) .

𝜕𝑧
Method 2: We perform partial differentiation. Recall the geometric significance of | (keep the remaining independent
𝜕𝑦 (𝑥 ,𝑦 )
0 0

variable 𝑥 fixed) that it is the slope of the tangent line to the intersecting curve 𝑧 = 𝑓(𝑥0 , 𝑦) at the point (𝑥0 , 𝑦0 , 𝑓(𝑥0 , 𝑦0 ))
on the cutting plane 𝑥 = 𝑥0 . In this example, the desired slope of the tangent line is given by

𝜕𝑧 𝜕 2
| = [𝑥 + 𝑦 2 ]| = 2𝑦|(𝑥=1 ,𝑦=2) = 4 .
𝜕𝑦 (𝑥=1 ,𝑦=2) 𝜕𝑦 (𝑥=1 ,𝑦=2)

Exercise 3:

Find all second-order partial derivatives of the functions below:

1. 𝑧 = 𝑓(𝑥, 𝑦) = sin(𝑥𝑦)
2. 𝑧 = 𝑓(𝑥, 𝑦) = ln |𝑥 + 𝑦|

Solution to Exercise 3:

𝜕2 𝑧 𝜕2 𝑧 𝜕2 𝑧 𝜕2 𝑧
There are four possible second-order partial derivatives listed as follows , , , . For the partial derivative
𝜕𝑥 2 𝜕𝑦 2 𝜕𝑥𝜕𝑦 𝜕𝑦𝜕𝑥
𝜕2 𝑧
, can you tell which operation is the first one to be performed?
𝜕𝑥𝜕𝑦

Part 1. Conclusion:

𝜕𝑧 𝜕2𝑧
𝑧 = sin(𝑥𝑦) = 𝑦 cos(𝑥𝑦) = −𝑦 2 sin(𝑥𝑦) 𝜕2𝑧 𝜕 𝜕𝑧
𝜕 𝜕𝑥 𝜕 𝜕𝑥 2 = ( ) = −𝑦 2 sin(𝑥𝑦)
[ ] [ ] 𝜕𝑥 2 𝜕𝑥 𝜕𝑥
𝜕𝑥 𝜕𝑥

𝜕𝑧 𝜕2𝑧
𝑧 = sin(𝑥𝑦) = 𝑥 cos(𝑥𝑦) = −𝑥 2 sin(𝑥𝑦) 𝜕2𝑧 𝜕 𝜕𝑧
𝜕 𝜕𝑦 𝜕 𝜕𝑦 2 = ( ) = −𝑥 2 sin(𝑥𝑦)
[ ] [ ] 𝜕𝑦 2 𝜕𝑦 𝜕𝑦
𝜕𝑦 𝜕𝑦

𝜕𝑧 𝜕2𝑧
𝑧 = sin(𝑥𝑦) = 𝑦 cos(𝑥𝑦) = cos(𝑥𝑦) − 𝑥𝑦 sin(𝑥𝑦) 𝜕2𝑧 𝜕 𝜕𝑧
𝜕 𝜕𝑥 𝜕 𝜕𝑦𝜕𝑥 = ( ) = cos(𝑥𝑦) − 𝑥𝑦 sin(𝑥𝑦)
[ ] [ ] 𝜕𝑦𝜕𝑥 𝜕𝑦 𝜕𝑥
𝜕𝑥 𝜕𝑦

𝜕𝑧 𝜕2𝑧
𝑧 = sin(𝑥𝑦) = 𝑥 cos(𝑥𝑦) = cos(𝑥𝑦) − 𝑥𝑦 sin(𝑥𝑦) 𝜕2𝑧 𝜕 𝜕𝑧
𝜕 𝜕𝑦 𝜕 𝜕𝑥𝜕𝑦 = ( ) = cos(𝑥𝑦) − 𝑥𝑦 sin(𝑥𝑦)
[ ] [ ] 𝜕𝑥𝜕𝑦 𝜕𝑥 𝜕𝑦
𝜕𝑦 𝜕𝑥

Practice Exercises 8 (edited by Dr. Hon-Ming HO)-page 3

Part 2. The four partial derivatives of 𝑧 = 𝑓(𝑥, 𝑦) = ln |𝑥 + 𝑦| are as follows:

𝜕2𝑧 −1 𝜕2𝑧 −1 𝜕2𝑧 𝜕2𝑧 −1

= , = , = = .
𝜕𝑥 2 (𝑥 + 𝑦)2 𝜕𝑦 2 (𝑥 + 𝑦)2 𝜕𝑦𝜕𝑥 𝜕𝑥𝜕𝑦 (𝑥 + 𝑦)2

Exercise 4:

We are given that equations 𝑢 = 𝑥 2 − 𝑦 2 and 𝑣 = 𝑥 2 − 𝑦 define 𝑥 and 𝑦 as functions of the independent variables 𝑢 and
𝜕𝑥 𝜕𝑦
𝑣 and the partial derivatives exist. Find , .
𝜕𝑢 𝜕𝑢

Solution to Exercise 4:

The underlying senses ( 基 礎 意 The underlying senses ( 基 礎 意 The underlying senses (基礎意
義) of objects 𝒙 and 𝒚 : 義) of objects 𝒖 and 𝒗 : 義 ) of the given pair of
equations 𝒖 = 𝒙𝟐 − 𝒚𝟐 and
𝒗 = 𝒙𝟐 − 𝒚 :

The objects 𝑥 and 𝑦 are treated The objects 𝑢 and 𝑣 are treated The given pair of equations is
as two independent variables as two dependent variables treated as a pair of working
whose values can freely vary (自 whose values are controlled and formulas for computing the
由 地 變 化 ) and independent of determined by 𝑥 and 𝑦 (𝑢 和 𝑣 values of 𝒖 and 𝒗 . Once the
each other. 之數值由 𝑥 和 𝑦 決定控制). values of 𝑥 and 𝑦 are given, the
working formulas enable us to
𝑥 𝑢 = 𝑥 2 − 𝑦2 compute the values of 𝑢 and
𝑦 𝑢
𝑣 .
𝑥 𝑣 = 𝑥2 − 𝑦
𝑣
𝑦

The second senses of objects 𝒙 The second senses of objects 𝒖 The second senses of the given
and 𝒚 : and 𝒗 : pair of equations 𝒖 = 𝒙𝟐 − 𝒚𝟐
and 𝒗 = 𝒙𝟐 − 𝒚 :

The objects 𝑥 and 𝑦 are no The objects 𝑢 and 𝑣 are no The given pair of equations is no
longer independent variables. To longer dependent variables. To longer treated as a pair of
𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 working formulas for computing
find , , 𝑥 and 𝑦 are find , , 𝑢 and 𝑣 are
𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢
𝑢 and 𝑣 . On the left hand side,
treated as two dependent treated as two independent
the first column, the numbers
variables whose values are variables controlling locally the
𝑢 , 𝑣 , 𝑥 and 𝑦 produced by
controlled by 𝑢 and 𝑣 . The values of 𝑥 and 𝑦 .
functions 𝑔1 and 𝑔2 are not
objects 𝑥 and 𝑦 are regarded as
randomly unrelated to each
two functions of 𝑢 and 𝑣 .
other. Instead, their relations are
𝑢 𝑥 = 𝑔1 (𝑢 , 𝑣) recorded by the following
𝑣 𝑔1
equations:
𝑢
𝑦 = 𝑔2 (𝑢 , 𝑣) 𝟐
𝑔2 𝒖 = (𝑔1 (𝑢 , 𝑣)) − (𝑔2 (𝑢 , 𝑣) )𝟐
𝑣
𝟐
𝒗 = (𝑔1 (𝑢 , 𝑣)) − 𝑔2 (𝑢 , 𝑣) .
In this case, we may find the
working formulas for 𝑔1 and 𝑔2 .
However, in other cases, their
formulas are not explicitly known.

Practice Exercises 8 (edited by Dr. Hon-Ming HO)-page 4

 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦
Now we are ready to find , . Bear in mind the following pair of equations when finding , .
𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢

𝟐 𝟐
𝒖 = (𝑔1 (𝑢 , 𝑣)) − (𝑔2 (𝑢 , 𝑣)) 𝒘𝒉𝒆𝒓𝒆 𝑥 = 𝑔1 (𝑢 , 𝑣) 𝑎𝑛𝑑 𝑦 = 𝑔2 (𝑢 , 𝑣) ,
𝟐
𝒗 = (𝑔1 (𝑢 , 𝑣)) − 𝑔2 (𝑢 , 𝑣) .

We perform implicit partial differentiation on both sides of the equations above with respect to variable 𝑢 . We obtain the
following equations

𝜕 𝜕 2 𝜕𝑥 𝜕𝑦 𝜕 𝜕 2 𝜕𝑥 𝜕𝑦
1= [𝑢] = [𝑥 − 𝑦 2 ] = 2𝑥 − 2𝑦 𝑎𝑛𝑑 0= [𝑣] = [𝑥 − 𝑦] = 2𝑥 − .
𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢
𝜕𝑥 𝜕𝑦
We obtain a system of two linear equations in unknowns , as follows.
𝜕𝑢 𝜕𝑢

𝜕𝑥 𝜕𝑦 𝟏 −𝟐𝒚 𝟐𝒙 𝟏
(𝟐𝒙) + (−𝟐𝒚) = 𝟏 𝜕𝑥 | | 1 𝜕𝑦 | | 1
{ 𝜕𝑢 𝜕𝑢 } ⟹ = 𝟎 −𝟏 = 𝑎𝑛𝑑 = 𝟐𝒙 𝟎 = .
𝜕𝑥 𝜕𝑦 𝜕𝑢 |𝟐𝒙 −𝟐𝒚| 2𝑥 − 4𝑥𝑦 𝜕𝑢 |𝟐𝒙 −𝟐𝒚| 1 − 2𝑦
(𝟐𝒙) + (− 𝟏 ) = 𝟎 𝟐𝒙 −𝟏 𝟐𝒙 −𝟏
𝜕𝑢 𝜕𝑢
𝜕𝑥 𝜕𝑦
The calculation above assumes that 𝑥 ≠ 0 and 1 − 2𝑦 ≠ 0 . If 𝑥 = 0 or 1 − 2𝑦 = 0 , then the partial derivatives , do
𝜕𝑢 𝜕𝑢
not exist. Why?

Practice Exercises 8 (edited by Dr. Hon-Ming HO)-page 5