Download Conditions for Conservative Vector Fields in Simply Connected Domains and more Study notes Calculus in PDF only on Docsity!
Math 241 - Calculus III
Spring 2012, section CL
§ 16.3. Conservative vector fields and simply connected domains
In these notes, we discuss the problem of knowing whether a vector field is conservative or not.
1 Conservative vector fields
Let us recall the basics on conservative vector fields.
Definition 1.1. Let
F : D → R
n be a vector field with domain D ⊆ R
n
. The vector field
F is said to be conservative if it is the gradient of a function. In other words, there is a
differentiable function f : D → R satisfying
F = ∇f. Such a function f is called a potential
function for
F.
Example 1.2.
F (x, y) = (x, y) is conservative, since it is
F = ∇f for the function f (x, y) =
1
2
(x
2
2
).
Example 1.3.
F (x, y) = (y cos x, sin x) is conservative, since it is
F = ∇f for the function
f (x, y) = y sin x.
The fundamental theorem of line integrals makes integrating conservative vector fields along
curves very easy. The following proposition explains in more detail what is nice about conser-
vative vector fields.
Proposition 1.4. The following properties of a vector field
F are equivalent.
F is conservative.
C
F · d~r is path-independent, meaning that it only depends on the endpoints of the curve
C.
C
F · d~r = 0 around any closed curve C.
Example 1.5. Find the line integral
C
F · d~r of the vector field
F (x, y) = (y cos x, sin x) along
the curve C parametrized by ~r(t) = (
π ln t
ln 16
, t
3
(^2) ), 1 ≤ t ≤ 4.
Solution. We know that
F is conservative, with potential function f (x, y) = y sin x. The
endpoints of C are ~r(1) = (0, 1) and ~r(4) = (
π ln 4
ln(
2 )
π
2
, 8). The fundamental theorem of
calculus yields
C
F · d~r =
C
∇f · d~r
= f (~r(4)) − f (~r(1))
= f (
π
, 8) − f (0, 1)
= 8 sin
π
− 1 sin 0
2 Necessary conditions
To know if a vector field
F is conservative, the first thing to check is the following criterion.
Proposition 2.1. Let D ⊆ R
2 be an open subset and let
F : D → R
2 be a continuously
differentiable vector field with domain D. If
F = (P, Q) is conservative, then it satisfies the
condition
P
y
= Q
x
everywhere on D.
Proof. Assume there is a differentiable function f : D → R satisfying
F = ∇f on D. Because f
is twice continuously differentiable (meaning it has all second partial derivatives f xx
, f xy
, f yx
, f yy
and they are all continuous), Clairaut’s theorem applies, meaning the mixed partial derivatives
agree. Since the first partial derivatives of f are (f x
, f y
) = (P, Q), we obtain
f xy
= f yx
(fx)y = (fy)x
P
y
= Q
x
Example 2.2. The vector field
F = (P, Q) = (ye
x , x
2 y
5 ) is not conservative, because the partial
derivatives are
P
y
= e
x
Q
x
= 2xy
5
6 = P y
Remark 2.3. Most vector fields are not conservative. If we pick functions P and Q “at random”,
then in general they will not satisfy P y
= Q
x
Remark 2.4. The analogue of proposition 2.1 in higher dimension R
n holds as well, since
Clairaut’s theorem holds in any dimension. However, there are more conditions to write down
in higher dimension, because there are more mixed second derivatives. Namely, there are
(
n
2
n(n−1)
2
conditions. In R
3
for example, a conservative vector field
F = (P, Q, R) satisfies
the 3 conditions
Py = Qx
P
z
= R
x
Q
z
= R
y
In §16.5, we will encode this information in something called the curl of
F.
Remark 2.5. Proposition 2.1 says that of
F = (P, Q) is conservative, then the function Q x
−P
y
is
identically zero. This function Qx − Py is sometimes called the (2-dimensional) curl of
F , which
will play an important role in §16.4. With that terminology, proposition 2.1 says “conservative
implies curl is zero”. Conveniently, that same statement will be the 3-dimensional analogue of
proposition 2.1 in §16.5.
Question 2.6. If a vector field
F = (P, Q) satisfies the condition P y
= Q
x
, is it automatically
conservative?
The velocity vector is
~r
′ (t) = (− sin t, cos t), 0 ≤ t ≤ 2 π.
and the integral of
F along C is
C
F · d~r =
2 π
0
F (~r(t)) · ~r
′
(t) dt
2 π
0
− sin t
cos
2 t + sin
2
t
cos t
cos
2 t + sin
2
t
) · (− sin t, cos t) dt
2 π
0
(− sin t, cos t) · (− sin t, cos t) dt
2 π
0
sin
2
t + cos
2
t dt
2 π
0
dt
= 2π 6 = 0.
Since we have found a closed curve along which the integral of
F is non-zero,
F cannot be
conservative.
3 Sufficient conditions
Depending on the shape of the domain D, the condition P y
= Q
x
is sometimes enough to
guarantee that the field is conservative.
Proposition 3.1. Let
F : R
2 → R
2 is a continuously differentiable vector field defined on R
2 .
If
F = (P, Q) satisfies the condition P y
= Q
x
, then
F is conservative.
Example 3.2. Consider the vector field
F = (y cos xy + 10x, x cos xy + 3y
2 ) defined on R
2 .
Determine whether
F is conservative, and if it is, find a potential function for it.
Solution. Writing P = y cos xy + 10x and Q = x cos xy + 3y
2
, we compute
P
y
= cos xy + y(−x sin xy)
= cos xy − xy sin xy
Q
x
= cos xy + x(−y sin xy)
= cos xy − xy sin xy
so that
F satisfies the condition P y
= Q
x
. Moreover, it is defined on all of R
2 , hence it is
conservative. Let us find a potential function f (x, y) for
F. We want
fx = P = y cos xy + 10x
f y
= Q = x cos xy + 3y
2
.
Using the first equation, we obtain
f =
P dx
y cos xy + 10x dx
= y
y
sin xy + 5x
2
= sin xy + 5x
2
whose derivative with respect to y is
x cos xy + g
′
(y).
Using the second equation, we equate this with Q:
x cos xy + g
′
(y) = x cos xy + 3y
2
⇒ g
′
(y) = 3y
2
⇒ g(y) = y
3
Choosing the constant c = 0, we obtain g(y) = y
3
and thus the potential function
f (x, y) = sin xy + 5x
2
3
.
Asking for
F to be defined (and continuously differentiable) on all of R
2
is somewhat restrictive.
In fact, that condition can be loosened, which gives the following improvement on proposition
Theorem 3.3. Let D ⊆ R
2 be open and simply connected, and let
F : D → R
2 is a continuously
differentiable vector field with domain D. If
F = (P, Q) satisfies the condition P y
= Q
x
, then
F is conservative (on D).
Theorem 3.3 did not apply to example 2.7, because the domain D = R
2
\ {(0, 0)} of
F was
not simply connected. Indeed, D is a punctured plane, and any curve going once around the
puncture cannot be retracted to its basepoint.
When theorem 3.3 does not apply because the domain is not simply connected, then we cannot
conclude from the condition P y
= Q
x
alone. A more subtle analysis is required.
Example 3.4. Consider the vector field
F =
1
x
2 +y
2
(x, y) on the domain
D = R
2
\ {(0, 0)}.
Is
F conservative?
Using the first equation, we obtain
f =
P dx
x
x
2
2
dx
u
du
substituting u = x
2
2
ln u + g(y) because u > 0 always
ln(x
2
2 ) + g(y) = ln
x
2
2
whose derivative with respect to y is
2 y
x
2
2
′
(y)
y
x
2
2
′
(y).
Using the second equation, we equate this with Q:
y
x
2
2
′
(y) =
y
x
2
2
⇒ g
′
(y) = 0
⇒ g(y) = c.
Choosing g(y) ≡ 0, we obtain a function
f (x, y) =
ln(x
2
2
)
which is indeed a potential for
F , defined on D. Therefore
F is conservative.
4 Restricting to a smaller domain
Does theorem 3.3 suggest that on a non simply connected domain, it is harder to find conser-
vative vector fields? No, there are still plenty of them.
Example 4.1. Let D = R
2 \ {(0, 0)} be the punctured plane and let
F : D → R
2 be the vector
field defined by
F (x, y) = (x, y). Then
F is conservative, since it is
F = ∇f for the function
f (x, y) =
1
2
(x
2
2 ) defined on D.
Example 4.2. Again, let D = R
2
\ {(0, 0)} and let
F : D → R
2
be the vector field defined
by
F (x, y) = (y cos x, sin x). Then
F is conservative, since it is
F = ∇f for the function
f (x, y) = y sin x defined on D.
More generally, pick any differentiable function f : D → R, then ∇f is a conservative vector
field on D. In particular, if we start with a nice function f : R
2
→ R such as f (x, y) = y sin x,
then ∇f is a conservative vector field on R
2 , as in examples 1.2 and 1.3. Moreover, its restriction
to any open subset D ⊆ R
2 is still conservative, with the same potential function (restricted to
D), as in examples 4.1 and 4.2.
Definition 4.3. Let X and Y be sets, f : X → Y a function, and S ⊆ a subset of X. The
restriction of f to S is the function f | S
: S → Y which is the same as f , but viewed as a
function defined on S only.
With this terminology, the comments above can be summarized as “the restriction of a conser-
vative vector field is conservative”.
We are not merely playing with words. Restricting to a smaller domain can change things
drastically.
Example 4.4. Consider the vector field
F =
1
x
2 +y
2
(−y, x) on the domain which is the “right
half-plane”
D = {(x, y) ∈ R
2
| x > 0 }.
Is
F conservative?
In 2.7, we have checked that
F satisfies P y
= Q
x
. Moreover, the domain D is simply connected.
Therefore, by theorem 3.3,
F is conservative.
Let us find a potential function f : D → R, which must satisfy
f x
= P =
−y
x
2
2
f y
= Q =
x
x
2
2
Using the second equation, we obtain
f =
Q dy
x
x
2
2
dy
x
x
2
y
x
2
) (^) dy which is fine because x 6 = 0
y
x
2
dy
x
= arctan
y
x
whose derivative with respect to x is
y
x
2
−y
x
2
′
(x)
−y
x
2
2
′
(x).
Using the first equation, we equate this with P :
−y
x
2
2
′
(x) =
−y
x
2
2
⇒ g
′
(x) = 0
⇒ g(x) = c.
