Geg 311 Vector Functions

GEG 311 Lecture _6: Multivariable functions analysis
Vector Functions
Partial, Directional Derivative
Differentiable Transformation
Gradients, Divergence, Curls Dr. Patrick S. Olayiwola
Dr. John Ogbemhe
Engr. Chidi Onyedikan
Vector Functions
Defn: Let 𝐴 ⊆ ℝ𝑛 be an open set. A function 𝑓: 𝐴 → ℝ𝑚 is
said to be differentiable at 𝑎 ∈ 𝐴 if there is aa linear
transformation, called the derivative of 𝑓 at 𝑎, 𝐷𝑎(𝑓):ℝ𝑛 →
ℝ𝑚 such that
𝑓 𝑥 − 𝑓 𝑎 − 𝐷𝑎(𝑓)(𝑥 − 𝑎)
lim
𝑥→𝑎 𝑥−𝑎
Denoting 𝑓 𝑎 + ℎ − 𝑓 𝑎 − 𝐷𝑎(𝑓)(𝑎)(ℎ) as 𝐸(ℎ), we
write
𝑓 𝑎 + ℎ − 𝑓 𝑎 = 𝐷𝑎 𝑓 ℎ + 𝐸(ℎ)
i.e., 𝑓: 𝐴 → ℝ𝑚 is said to be differentiable at 𝑎 ∈ 𝐴 if there
is a linear transformation 𝐷𝑎(𝑓)(ℎ) such that
𝑓 𝑎 + ℎ − 𝑓 𝑎 = 𝐷𝑎 𝑓 ℎ + 𝐸(ℎ)
𝐸(𝑥−𝑎)
With 𝐸(ℎ) a function that satisfies lim =0
ℎ→0 ℎ
Vector Functions
Example: Let
ℝ3 × ℝ3 → ℝ
𝑓:
(𝒙, 𝒚) → 𝒙 ∙ 𝒚
be the dot product in ℝ3. Show that 𝑓 is differentiable and
that
𝐷(𝒙,𝒚) 𝑓 𝒉, 𝒌 = 𝒙 ∙ 𝒚 + 𝒉 ∙ 𝒌
Solution: 𝑓 𝒙 + 𝒉, 𝒚 + 𝒌 − 𝑓 𝒙, 𝒚 = 𝒙 + 𝒉 ∙ 𝒚 + 𝒌
=𝒙∙𝒚+𝒙∙𝒌+𝒉∙𝒚+𝒉∙𝒌−𝒙∙𝒚
=𝒙∙𝒌+𝒉∙𝒚+𝒉∙𝒌
Since 𝒙 ∙ 𝒌 + 𝒉 ∙ 𝒚 is a linear function of (𝒉, 𝒌) if we choose
𝑬 𝒉 = 𝒉 ∙ 𝒌, using CBS inequality, i.e., 𝒉 ∙ 𝒌 ≤ 𝒉 𝒌 and
𝑬(𝒉)
lim ≤ 𝒌 = 0 QED
(𝑥,𝑦)→(0,0) 𝒉
Partial and Directional Derivative
Defn: Let 𝐴 ⊆ ℝ𝑛 , a function 𝑓: 𝐴 → ℝ𝑚 and put
𝑓1(𝑥1, … , 𝑥𝑛)
𝑓2(𝑥1, … , 𝑥𝑛)
𝑓 𝒙 =
⋮
𝑓𝑚(𝑥1, … , 𝑥𝑛)
𝜕𝑓𝑖
Here 𝑓𝑖 : ℝ𝑛 → ℝ. The partial derivative is defined as
𝜕𝑥𝑗
𝜕𝑓𝑖 𝑓𝑖 𝑥1,…,𝑥𝑗+ℎ,…,𝑥𝑛 −𝑓𝑖(𝑥1,…,𝑥𝑗,…,𝑥𝑛)
𝜕𝑗 𝑓𝑖 = =lim
𝜕𝑥𝑗 ℎ→0 ℎ
whenever the limit exist.
Example: If 𝑓: ℝ3 → ℝ and 𝑓 𝑥, 𝑦, 𝑧 = 𝑥 + 𝑦2 + 𝑧3 + 3𝑥𝑦2𝑧3, then
𝜕𝑓
𝑥, 𝑦, 𝑧 = 1 + 3𝑦2𝑧3,
𝜕𝑥
𝜕𝑓
𝑥, 𝑦, 𝑧 = 2𝑦 + 6𝑥𝑦𝑧3,
𝜕𝑦
𝜕𝑓
𝑥, 𝑦, 𝑧 = 3𝑧2 + 9𝑥𝑦2𝑧2,
𝜕𝑧
Partial and Directional Derivative
Let 𝑓(𝒙) be a vector-valued function. Then its derivative in
the direction 𝒖 is defined as
𝑑
𝐷𝑢𝑓 𝒙 ≔ 𝐷𝑓 𝒙 𝒖 = 𝑓 𝒗 + 𝛼𝒖 𝛼 = 0,
𝑑𝛼
for all values of 𝒖.
Differentiable Transformations
Let 𝐴 ⊆ ℝ𝑛, 𝐵 ⊆ ℝ𝑚 be open sets 𝒇, 𝒈: 𝐴 → ℝ𝑚, 𝛼 ∈ ℝ, be
differentiable on 𝐵, and 𝑓(𝐴) ⊆ 𝐵. Then we have
Addition rule: 𝐷𝒙 𝒇 + 𝛼𝒈 = 𝐷𝒙 𝒇 + 𝐷𝒙 𝒈
Chain rule: 𝐷𝒙 𝒉 ⋄ 𝒇 = 𝐷𝒇(𝒙) 𝒉 + 𝐷𝒙 𝒇
Example: Let
𝒇 𝑢, 𝑣 = (𝑢𝑒𝑣, 𝑢 + 𝑣, 𝑢𝑣)
𝒉 𝑥, 𝑦 = (𝑥2 + 𝑦, 𝑦 + 𝑧)
Find 𝒇 ⋄ 𝒉 ’ 𝑥, 𝑦 .
Solution:
𝑒𝑣 𝑢𝑒𝑣
2𝑥 1 0
𝒇 𝑢, 𝑣 = 1 1 and ℎ′ 𝑥, 𝑦 =
𝑣 𝑢 0 1 1
Observe that
𝑒 𝑦+𝑧 (𝑥2 + 𝑦)𝑒 𝑦+𝑧
𝒇′ ℎ 𝑥, 𝑦 = 1 1
𝑦+𝑧 𝑥2 + 𝑦
𝑓 ⋄ ℎ ’ 𝑥, 𝑦 = 𝑓 ′ ℎ 𝑥, 𝑦 ℎ′ 𝑥, 𝑦
𝑒 𝑦+𝑧 (𝑥2 + 𝑦)𝑒 𝑦+𝑧
2𝑥 1 0
= 1 1
2 0 1 1
𝑦+𝑧 𝑥 +𝑦
2𝑥𝑒 𝑦+𝑧 (1 + 𝑥2 + 𝑦)𝑒 𝑦+𝑧 (𝑥2 + 𝑦)𝑒 𝑦+𝑧

= 2𝑥 2 1
2𝑥𝑦 + 2𝑥𝑧 𝑥2 + 2𝑦 + 𝑧 𝑥2 + 𝑦
Example: Let
𝑓: ℝ2 → ℝ, 𝒇 𝑢, 𝑣 = u2 + ev
⇒ 𝑢, 𝑣: ℝ3 → ℝ ⇒ 𝑢 𝑥, 𝑦 = 𝑦 + 𝑧
Put 𝑢 𝑥, 𝑦 = 𝑓(𝑢 𝑥, 𝑦, 𝑧 , 𝑣 𝑥, 𝑦, 𝑧 . Find the partial
derivatives of ℎ.
Solution: Put 𝑔: ℝ3 → ℝ2, 𝑔 𝑥, 𝑦 = (𝑢 𝑥, 𝑦 , 𝑣 𝑥, 𝑦 =
(𝑥𝑧, 𝑦 + 𝑧). Observe that h = 𝑓 ⋄ 𝑔. Now,
′ 𝑧 0 𝑥
𝒈 (𝑥, 𝑦) =
0 1 1
𝒇′(𝑢, 𝑣) = 2𝑢 𝑒𝑣
𝒇′(ℎ(𝑥, 𝑦)) = 2𝑥𝑧 𝑒 𝑦+𝑧
Thus,
𝜕ℎ 𝜕ℎ 𝜕ℎ
𝑥, 𝑦 𝑥, 𝑦 (𝑥, 𝑦) = ℎ′ 𝑥, 𝑦
𝜕𝑥 𝜕𝑦 𝜕𝑧
= (𝒇′ 𝒈 𝑥, 𝑦 )(𝒈′ 𝑥, 𝑦 )
𝑧 0 𝑥
= 2𝑥𝑧 𝑒 𝑦+𝑧 = 2𝑥𝑧2 𝑒 𝑦+𝑧 2𝑥2𝑧 + 𝑒 𝑦+𝑧
0 1 1
Equating into components
𝜕ℎ
𝑥, 𝑦 = 2𝑥𝑧2,
𝜕𝑥
𝜕ℎ
𝑥, 𝑦 = 𝑒 𝑦+𝑧
𝜕𝑦
𝜕ℎ
𝑥, 𝑦 = 2𝑥2𝑧 + 𝑒 𝑦+𝑧
𝜕𝑧
Some Defns: A function 𝑓 is said to be continuously differentiable if the derivative 𝑓’

exists and is itself a continuous function.
Continuously differentiable functions are said to be of class 𝐶 1 . A function is of class
𝐶 2 if the first and second derivatives of the function exist and continuous. More
generally, a function is said to be of class 𝐶 𝑘 if the first k derivatives exist and
( )
𝑛
continuous. If derivatives 𝑓 exist for all positive integers n, the function is said to be
smooth or of class 𝐶 ∞ .
Gradients, Curls & Directional Derivatives
Defn: Let
ℝ𝑛 → ℝ
𝑓:
𝒙 → 𝑓(𝒙)
be a scalar field. The gradient of 𝑓 is the vector defined and
denoted by
∇𝑓 𝒙 ≔ 𝐷𝑓 𝒙 ≔ 𝜕1𝑓 𝒙 , 𝜕2𝑓 𝒙 , … , 𝜕𝑛𝑓(𝒙).
The gradient operator is
∇= (𝜕1,𝜕2,…, 𝜕𝑛).
Theorem: Let 𝐴 ⊆ ℝ𝑛 be open and let 𝑓: 𝐴 → ℝ be a scalar

field, and that 𝑓 is differentiable in 𝐴. Let 𝐾 ∈ ℝ be a
constant. Then ∇𝑓 𝒙 is orthogonal to the surface implicitly
defined by 𝑓 𝒙 = 𝐾.
Proof: Let
ℝ → ℝ𝑛
𝒄:
𝑡 → 𝒄(𝑡)
be a curve lying on this surface. Choose 𝑡0 so that
𝐜 𝑡0 = 𝒙 . Then
𝑓 ∘ 𝒄 𝑡0 = 𝑓 𝒄 𝑡 = 𝐾
and using the chain rule
D𝑓 𝑐 𝑡0 D𝑐 𝑡0 = 0
which translates to
(∇𝑓 𝒙 ) ∙ (𝒄′ 𝑡0) = 0
Since 𝒄′ 𝑡0 is tangent to the surface and its dot product
with ∇𝑓 𝒙 is zero, it is then concluded that ∇𝑓 𝒙 is normal
to the surface.
Example 1: find a unit vector normal to the surface 𝑥3 +
𝑦3 + z=4 at the point 1,1,2 .
Solution: 𝑓 𝑥, 𝑦, 𝑧 = 𝑥3 + 𝑦3 + 𝑧 − 4 has gradient
∇𝑓 𝑥, 𝑦, 𝑧 = (3𝑥2,3𝑦2,1)
which at 1,1,2 is 3,3,1 . Normalizing gives
3 3 1
, ,
19 19 19
Example 2: Find the direction of the greatest rate of
increase of 𝑓 𝑥, 𝑦, 𝑧 = 𝑥𝑦𝑒𝑧 t the point 2,1,2 .
Solution: The direction is that of the gradient vector.
∇𝑓 𝑥, 𝑦, 𝑧 = (𝑦𝑒𝑧, 𝑥𝑒𝑧, 𝑥𝑦𝑒𝑧)
So at 2,1,2 , we have (𝑒2,2𝑒2,2𝑒2). Normalizing to give
1
1,1,2 .
5
Example 3: Sketch the gradient vector field for
𝑓 𝑥, 𝑦 = 𝑥2 + 𝑦2
As well as several contours for this function.
Solution: The contours for this function are the curves
defined by
𝑓 𝑥, 𝑦 = 𝑘
for various values of 𝑘. So,
𝑥2 + 𝑦2 = 𝑘
which gives circles centred at the origin with radius 𝑘.
The gradient vector field is
∇𝑓 𝑥, 𝑦 = 2𝑥𝒊 + 2𝑦𝒋
𝑧 2
Example 4: If 𝑓 𝑥, 𝑦, 𝑧 = 2 2
(𝑥 , 𝑦 , 𝑦𝑒 )
then
𝑧 2
𝑑𝑖𝑣𝑓 𝒙 = (2𝑥, 2𝑦, 2𝑦𝑧𝑒 )
Mean Value Theorem for Scalar Fields
Let 𝑈 be an open connected subset of ℝ𝑛 and let 𝑓: 𝑈 → ℝ
be a differentiable function. Fix points 𝒙, 𝒚 ∈ 𝑈 such that the
segment connecting 𝒙 to 𝒚 lies in 𝑈. Then,
𝑓 𝒚 − 𝑓 𝒙 = ∇𝑓(𝒛) ∙ (𝒚 − 𝒙)
Where 𝒛 is a point in the open segment connecting 𝒙 and 𝒚.
Proof: Let 𝑈 be an open connected subset of ℝ𝑛. Now define

𝑔 𝑡 ≔ 𝑓 1 − 𝑡 𝒙 + 𝑡𝒚. Since 𝑓 is a differentiable function
in 𝑈 the function 𝑔 is a continuous function in [0,1] and
differentiable in (0,1).
The mean value theorem give
𝑔 1 − 𝑔 0 = 𝑔′(𝑐)
For some 𝑐 ∈ (0,1). But since 𝑔 0 = 𝑓(𝒙) and 𝑔 1 =
𝑓(𝒚), we have
𝑓 𝒚 − 𝑓 𝒙 = ∇𝑓 1 − 𝑐 𝒙 + 𝑐𝒚 ⋅ (𝒚 − 𝒙)
or 𝑓 𝒚 − 𝑓 𝒙 = ∇𝑓 𝒛 ⋅ (𝒚 − 𝒙)
where 𝒛 is a point in the open segment connecting 𝒙 and 𝒚.
By the CBS inequality.
𝑓 𝒚 − 𝑓 𝒙 = ∇𝑓 1 − 𝑐 𝒙 + 𝑐𝒚 𝒚 − 𝒙
Curl: If 𝐹: ℝ3 → ℝ3 is a vector field with components 𝑭 =
(𝐹1, 𝐹2, 𝐹3). The curl of 𝑭 is defined as
𝜕2𝐹3 − 𝜕3𝐹2
∇ × 𝑭 = 𝜕3𝐹1 − 𝜕1𝐹3
𝜕1𝐹2 − 𝜕2𝐹1
In its n-dimensions: Let 𝑔𝑘: ℝn → ℝ𝑛, 1 ≤ 𝑘 ≤ 𝑛 − 2 be

vector fields with 𝑔𝑖 = (𝑔𝑖1, 𝑔𝑖2, … , 𝑔𝑖𝑛). Then, the curl of
𝑔1, 𝑔2, … , 𝑔𝑛 − 2
𝑐𝑢𝑟𝑙 𝑔1, 𝑔2, … , 𝑔𝑛 − 2 𝒙
𝑒1 𝑒2 … 𝑒𝑛
𝜕1 𝜕2 … 𝜕𝑛
= 𝑑𝑒𝑡 𝑔11(𝒙) 𝑔12(𝒙) … 𝑔1𝑛(𝒙)
⋮ ⋮ ⋮ ⋮
𝑔𝑛 − 2,1(𝒙) 𝑔𝑛 − 2,2(𝒙) … 𝑔𝑛 2 𝑛 𝒙)
− , (
Example: If 𝑓 𝑥, 𝑦, 𝑧, 𝑤 = 𝑒 𝑥𝑦𝑧 , 0,0, 𝑤 2 , 𝑔 𝑥, 𝑦, 𝑧, 𝑤 =
0,0, 𝑧, 0 , then
𝑐𝑢𝑟𝑙 𝑓, 𝑔 𝑥, 𝑦, 𝑧, 𝑤
𝒆𝟏 𝒆𝟐 𝒆𝟑 𝒆𝟒
𝜕1 𝜕2 𝜕3 𝜕4
= 𝑑𝑒𝑡 𝑒 𝑥𝑦𝑧 0 0 𝑤2 = 𝑥𝑧 2 𝑒 𝑥𝑦𝑧 𝒆4
0 0 𝑧 0
Defn: Let 𝐴 ⊆ ℝ𝑛 be open and let 𝑓: 𝐴 → ℝ be a scalar

field, and assume that 𝑓 is differentiable in 𝐴. Let 𝒗 ∈
ℝ𝑛\ 𝟎 be such that 𝐱 + t𝒗 ∈ 𝐴 for sufficiently small 𝑡 ∈ ℝ.
Then, the directional derivative of 𝒇 in the direction 𝒗 at
the point 𝒙 is given by ∇𝑓 𝑥 ∙ 𝒗
Example: Find the directional derivative of
𝑓 𝑥, 𝑦, 𝑧 = 𝑥 3 + 𝑦 3 − 𝑧 2 in the direction of (1,2,3).
Solution: we have
∇𝑓 𝑥, 𝑦, 𝑧 = (3𝑥 2 , 3𝑦 2 , −2𝑧)
and,
∇𝑓 𝑥, 𝑦, 𝑧 ∙ 𝒗 = 3𝑥 2 + 6𝑦 2 − 6𝑧.
Theorems or Rules in Differentiation
Theorem:
∇ ∙ 𝜓𝒖 = 𝜓∇ ∙ 𝒖 + 𝒖 ∙ ∇𝜓
∇ × 𝜓𝒖 = 𝜓∇ × 𝒖 + 𝒖 × ∇𝜓
∇∙𝒖×𝒗=𝒗∙∇×𝒖−𝒖∙∇×𝒗
∇ × 𝒖 × 𝒗 = 𝒗 ∙ ∇𝒖 − 𝒖 ∙ ∇𝒗 + 𝒖 ∇ ∙ 𝒗 − 𝒗(∇ ∙ 𝒖)
∇ 𝒖 ∙ 𝒗 = 𝒖 ∙ ∇𝒗 + 𝒗 ∙ ∇𝒖 + 𝒖 × ∇ × 𝒗 + 𝒗 × (∇ × 𝒖)
∇ × ∇𝜓 = 𝑐𝑢𝑟𝑙 𝑔𝑟𝑎𝑑𝜓 = 0
∇ ∙ ∇ × 𝒖 = 𝑑𝑖𝑣 𝑐𝑢𝑟𝑙𝒖 = 0
∇ ∙ ∇𝜓1 × ∇𝜓2 = 0
∇ × ∇ × 𝒖 = 𝑐𝑢𝑟𝑙(𝑐𝑢𝑟𝑙u)=∇ ∇ ∙ 𝒖 − ∇2𝒖

Geg 311 Vector Functions

Uploaded by

Copyright:

Available Formats

Geg 311 Vector Functions

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Geg 311 Vector Functions

Uploaded by

Copyright:

Available Formats

GEG 311 Lecture _6: Multivariable functions analysis

2𝑥𝑒 𝑦+𝑧 (1 + 𝑥2 + 𝑦)𝑒 𝑦+𝑧 (𝑥2 + 𝑦)𝑒 𝑦+𝑧

Some Defns: A function 𝑓 is said to be continuously differentiable if the derivative 𝑓’

Theorem: Let 𝐴 ⊆ ℝ𝑛 be open and let 𝑓: 𝐴 → ℝ be a scalar

Proof: Let 𝑈 be an open connected subset of ℝ𝑛. Now define

In its n-dimensions: Let 𝑔𝑘: ℝn → ℝ𝑛, 1 ≤ 𝑘 ≤ 𝑛 − 2 be

Defn: Let 𝐴 ⊆ ℝ𝑛 be open and let 𝑓: 𝐴 → ℝ be a scalar

You might also like