Math 2011-Introduction To Multivariable Calculus (Edited by Dr. Hon-Ming HO) Practice Exercises 10: Gradient Vectors and Directional Derivatives
Math 2011-Introduction To Multivariable Calculus (Edited by Dr. Hon-Ming HO) Practice Exercises 10: Gradient Vectors and Directional Derivatives
Math 2011-Introduction To Multivariable Calculus (Edited by Dr. Hon-Ming HO) Practice Exercises 10: Gradient Vectors and Directional Derivatives
Hon-Ming HO)
Practice Exercises 10: Gradient Vectors and Directional Derivatives
Exercise 1:
Find the derivative of the functions below at point 𝑃0 (𝑥0 , 𝑦0 ) in the specified direction of 𝑣⃑
Solution to Exercise 1:
We are required to find directional derivative (𝐷𝑢⃑⃑ 𝑓)(𝑃0 ) at the specific point in the given direction. There are two ways to find
the desired directional derivatives for three parts above.
Assume that ⃑⃑ is a unit vector. Recall the limit definitions of directional derivatives.
𝒖
𝑓(𝑥0 + ℎ𝑢1 , 𝑦0 + ℎ𝑢2 ) − 𝑓(𝑥0 , 𝑦0 )
2 − variable version: (𝐷𝑢⃑⃑ 𝑓)(𝑃0 ) = (𝐷<𝑢1 ,𝑢2> 𝑓)(𝑥0 , 𝑦0 ) = lim .
ℎ→0 ℎ
𝑓(𝑥0 + ℎ𝑢1 , 𝑦0 + ℎ𝑢2 , 𝑧0 + ℎ𝑢3 ) − 𝑓(𝑥0 , 𝑦0 , 𝑧0 )
3 − variable version: (𝐷𝑢⃑⃑ 𝑓)(𝑃0 ) = (𝐷<𝑢1 ,𝑢2 ,𝑢3> 𝑓)(𝑥0 , 𝑦0 , 𝑧0 ) = lim .
ℎ→0 ℎ
(𝐷𝑢⃑ 𝑓)(𝑃0 ) = 𝑢
⃑ ⦁ (𝛻𝑓)(𝑃0 ) .
? ? ?
𝒄𝒐𝒔−𝟏 ῧ ? ? 𝐜𝐨𝐬 ῧ (General cosine rule)
(▭1 ) ∙ (▭2 ) ?
𝐭𝐚𝐧−𝟏 ῧ 𝐭𝐚𝐧 ῧ
(× Rule)
𝑥𝑦 𝑥𝑦
We observe that 𝑠𝑖𝑛 −1 ( ) = 𝒔𝒊𝒏−𝟏 ῧ where ῧ = . Then we apply the general arcsine rule for partial differentiation as
2 2
follows:
𝑦
𝜕 𝑥𝑦 𝜕 1 𝜕 1 𝜕 𝑥𝑦 ( )
[𝑠𝑖𝑛−1 ( )] = [𝒔𝒊𝒏−𝟏 ῧ ] = ∙ [ῧ]= ∙ [ ]= 2 .
𝜕𝑥 2 𝜕𝑥 2 𝜕𝑥 𝑥𝑦 2 𝜕𝑥 2 𝑥𝑦 2
√1 − ( ῧ ) √1 − ( ) √1 − ( )
2 2
Exercise 2:
Find the unit vectors along which the given functions below increase and decrease most rapidly at 𝑃0 . Then find the derivatives
of the functions in these directions.
Solution to Exercise 2:
𝜕𝑓 𝜕𝑓
(𝛻𝑓)(𝑃0 ) = < , >| = (2𝑥𝑦 + 𝑦𝑒 𝑥𝑦 sin 𝑦) 𝑖⃑ + (𝑥 2 + 𝑥𝑒 𝑥𝑦 sin 𝑦 + 𝑒 𝑥𝑦 cos 𝑦)𝑗⃑]𝑥=1 ,𝑦=0 = 2𝑗⃑ .
𝜕𝑥 𝜕𝑦 𝑥=1 ,𝑦=0
(𝛻𝑓)(𝑃0 )
(𝐷𝑢⃑ 𝑓)(𝑃0 ) = 𝑢
⃑ ⦁ (𝛻𝑓)(𝑃0 ) = ⦁ (𝛻𝑓)(𝑃0 ) = |(𝛻𝑓)(𝑃0 )| = 2 ,
|(𝛻𝑓)(𝑃0 )|
−(𝛻𝑓)(𝑃0 )
(𝐷−𝑢⃑ 𝑓)(𝑃0 ) = (−𝑢
⃑ )⦁ (𝛻𝑓)(𝑃0 ) = ⦁ (𝛻𝑓)(𝑃0 ) = −|(𝛻𝑓)(𝑃0 )| = −2 .
|(𝛻𝑓)(𝑃0 )|
(𝛻𝑓)(𝑃 ) 1 1 1 −(𝛻𝑓)(𝑃0 ) −1 −1 −1
⃑⃑ = |(𝛻𝑓)(𝑃0)| =<
Part B. The desired unit vectors are 𝑢 , , > and −𝑢
⃑⃑ = |(𝛻𝑓)(𝑃0 )|
=< , , > . Moreover, we have
0 √3 √3 √3 √3 √3 √3
Exercise 3:
Solution to Exercise 3:
There are more than one method to find an equation of the desired tangent line at the stated point 𝑃 . We once learnt the first
(non-vectorial) method in Calculus I (math 1013), using the technique of implicit differentiation. We are going to present a vectorial
method. How do we enter into vectorial perspective?
1) Things to do: Find a normal vector to the given plane curve. To find such 𝑦
a normal vector, we treat the given plane curve as a level curve of some
function 𝑧 = 𝑓(𝑥, 𝑦) at a certain height 𝑧 = 𝑐 . In this example, there
are infinitely many possible choices of the target function 𝑓 . One of 𝑃(−𝟏 , 𝟐)
the natural choices is 𝑧 = 𝑓(𝑥, 𝑦) = 𝑥 2 − 𝑥𝑦 + 𝑦 2 at height 𝑧 = 7 .
We know that the gradient vector 𝛻𝑓(𝑥, 𝑦) is perpendicular to the
𝑄(𝑥, 𝑦) 𝑥
level curve of 𝑧 = 𝑓(𝑥, 𝑦) that passes through point (−𝟏 , 𝟐) . Since
𝜕𝑓 𝜕𝑓
𝛻𝑓(𝑥, 𝑦) =< , >=< 2𝑥 − 𝑦 , 2𝑦 − 𝑥 > , we have
𝜕𝑥 𝜕𝑦
𝜵𝒇(−𝟏 , 𝟐) =< −𝟒 , 𝟓 > which is normal to the level curve (ellipse).
2) Things to do: Use the obtained gradient vector above to construct the equation of the required tangent line. Take a testing
point (𝑥, 𝑦) on the tangent line. The vector from the point of tangency to the testing point must be perpendicular to the
gradient vector. In this example, we let 𝑄(𝑥, 𝑦) be a testing point on the tangent line and 𝑃(−𝟏 , 𝟐) the point of tangency.
Thus we have
⃑⃑⃑⃑⃑
𝑃𝑄 ⦁ 𝜵𝒇(−𝟏 , 𝟐) = 0 ⇔ < 𝑥 − (−𝟏) , 𝑦 − 𝟐 > ⦁ < −𝟒 , 𝟓 > = 0
Conclusion: The equation of the required tangent line at point (−1 ,2) is given by −4𝑥 + 5𝑦 − 14 = 0 .
Remark: The second vectorial method can be generalized to find the equation of the tangent plane to a surface at a particular
point.
Exercise 4:
The derivative of 𝑧 = 𝑓(𝑥, 𝑦) at point 𝑃0 (1 ,2) in direction of 𝑖⃑ + 𝑗⃑ is equal to 2√2 and in the direction of −2𝑗⃑ is equal to
−3 . What is the derivative of 𝑧 = 𝑓(𝑥, 𝑦) at the point 𝑃0 (1 ,2) in the direction of −𝑖⃑ − 2𝑗⃑ ? Show your steps.
Solution to Exercise 4:
𝑖⃑+𝑗⃑ 1 1 −2𝑗⃑
The question above is equivalent to the following statement. Let ⃑⃑⃑⃑⃑
𝑢1 = = 𝑖⃑ + 𝑗⃑ and ⃑⃑⃑⃑⃑
𝑢2 = = −𝑗⃑ and ⃑⃑⃑⃑⃑
𝑢3 =
√12 +12 √2 √2 √02 +(−2)2
−𝑖⃑−2𝑗⃑ −1 −2
= 𝑖⃑ + 𝑗⃑ . If (𝐷⃑⃑⃑⃑⃑
𝑢1 𝑓)(1 ,2) = 2√2 and (𝐷⃑⃑⃑⃑⃑
𝑢2 𝑓)(1 ,2) = −3 , then what is the value of the directional
√(−1)2 +(−2)2 √5 √5
derivative (𝐷⃑⃑⃑⃑⃑
𝑢3 𝑓)(1 ,2) ?
(𝐷𝑢⃑ 𝑓)(𝑃0 ) = 𝑢
⃑ ⦁ (𝛻𝑓)(𝑃0 ) .
The underlying sense of (𝐷𝑢⃑ 𝑓)(𝑃0 ) : The vectorial sense (向量性的意義) of (𝐷𝑢⃑ 𝑓)(𝑃0 ) :
(𝐷𝑢⃑ 𝑓)(𝑃0 ) is the rate of change of the value of 𝑓 at (𝐷𝑢⃑ 𝑓)(𝑃0 ) can be realized as a dot product of the
the point 𝑃0 along the direction of the unit vector 𝑢
⃑⃑ . direction vector 𝑢
⃑⃑ and the gradient vector of 𝑓 .
1 𝜕𝑓 1 𝜕𝑓
2√2 = (𝐷⃑⃑⃑⃑⃑
𝑢1 𝑓)(1 ,2) = ∙ (1,2) + ∙ (1,2)
√2 𝜕𝑥 √2 𝜕𝑦
𝜕𝑓 𝜕𝑓
−3 = (𝐷⃑⃑⃑⃑⃑
𝑢2 𝑓)(1 ,2) = 0∙ (1,2) + (−1) (1,2)
𝜕𝑥 𝜕𝑦
−1 𝜕𝑓 −2 𝜕𝑓
? = (𝐷⃑⃑⃑⃑⃑
𝑢3 𝑓)(1 ,2) = ∙ (1,2) + ∙ (1,2)
√5 𝜕𝑥 √5 𝜕𝑦
𝜕𝑓 𝜕𝑓 𝜕𝑓
Solve the first two equations for the unknown partial derivatives (1,2) and (1,2) . We obtain that (1,2) = 1 and
𝜕𝑥 𝜕𝑦 𝜕𝑥
𝜕𝑓 −7
(1,2) = 3 . Plug them into the 3rd equation, we have (𝐷⃑⃑⃑⃑⃑
𝑢3 𝑓)(1 ,2) = .
𝜕𝑦 √5