Partial Differential Equations III/IV
[6pt] Exercise Sheet 4: Solutions
-
1.
Green’s functions. By the Fundamental Theorem of Calculus, integrating over , for any , gives
where we have used the boundary condition . Integrating again, this time over , gives
Taking and using the boundary condition yields
Therefore
By interchanging the order of integration we can write this as
Therefore
with
-
2.
Homogenization.
-
(i)
Integrate over :
Now integrate over :
(1) We determine by evaluating this expression at :
Substituting this into (1) gives
as required.
-
(ii)
Taking gives
We are told in the hint to use the Riemann-Lebesgue Lemma, which states that if is –periodic, then for any interval ,
(2) Applying (2) with , , , and on gives
(Technical remark: We cannot take for all , else . But we can take to be any function in such that on . The choice of outside does not matter since it does not affect the integrals in (2).)
-
(iii)
This is simply a matter of interchanging the order of integration:
with
-
(iv)
Clearly satisfies the boundary conditions. By the Fundamental Theorem of Calculus, differentiating equation (3) gives
Differentiating again gives
Therefore
as required.
-
(v)
By definition,
On the other hand,
Therefore . In fact the Cauchy-Schwarz inequality can be used to show that
for any choice of .
-
(vi)
Without loss of generality we can assume that . Using the hint and integration by parts gives
(4) Let and let denote floor, which is the largest integer less than or equal to . Since is 1–periodic,
(5) Observe that
Therefore by the Pinching Lemma (Squeezing Lemma)
(6) Also
Therefore
(7) By combining equations (5), (6), (7) we find that
(8) Therefore the limit of the first term on the right-hand side of equation (4) is
(9)
-
(i)
-
3.
Radial symmetry of Laplace’s equation on . Let be a harmonic function. Let and define by . Then
To be precise
(This can also be written as .)
-
4.
Fundamental solution of Poisson’s equation in 3D.
-
(i)
One way of computing is using spherical polar coordinates:
Another way of computing is as follows:
-
(ii)
Let be compact. Since is bounded, there exists such that . Therefore
Therefore .
-
(iii)
By part (i),
Therefore .
-
(iv)
By the Chain Rule
Let be compact. Since is bounded, there exists such that . Therefore
Therefore .
-
(i)
-
5.
Fundamental solution of Poisson’s equation in 1D. We compute
(symmetry of convolution) (integration by parts) (Fundamental Theorem of Calculus) as required.
-
6.
The function spaces and . Let , , . By integrating we see that
-
(i)
for ,
-
(ii)
for ,
-
(iii)
for ,
-
(iv)
for any (by parts (i),(ii)).
-
(i)
-
7.
Properties of the convolution.
-
(i)
Let , and let . Choose such that . In particular, outside the interval . Therefore
since and is compact.
-
(ii)
Now assume that . By Lemma 4.12, . Therefore
Therefore
and so , as required.
-
(iii)
The convolution is commutative since
as required.
-
(i)
-
8.
The Poincaré inequality for functions that vanish on the boundary. Let satisfy . Then
since . Therefore
(Cauchy-Schwarz) Squaring and integrating gives
This is the Poincaré inequality with .
-
9.
The Poincaré inequality on unbounded domains.
-
(i)
For , define by
(Exercise: Sketch to get a better understanding of the example.) Observe that
Therefore . We also have since
We compute
which is independent of . But
Therefore
as required. This means that, given any , we can choose large enough so that
which means that the Poincaré inequality on does not hold. We constructed this counter example using spreading; the support of spreads as without changing the –norm of .
-
(ii)
Let . Let with and with for all . Then
as required.
-
(i)
-
10.
The Poincaré constant depends on the domain. There exits such that
(12) for all with . Let with . Then
with , as desired.
-
11.
Eigenvalues of : Can you hear the shape of a drum? Multiply the PDE by (the complex conjugate of ) and integrate over :
The boundary condition on implies that on and so
as required.
-
12.
The optimal Poincaré constant and eigenvalues of .
-
(i)
Multiply the PDE by and integrate over :
since on . Rearranging gives
-
(ii)
Let minimise . Let . Define and define , . Since is minimised by , then is minimised by . It follows that
The numerator must be zero. Rearranging gives
Integrating by parts gives
Since this holds for all , the Fundamental Lemma of the Calculus of Variations implies that
If we define
then
In other words, is an eigenfunction of . By definition
Since minimises , then must be the smallest eigenvalue of on , i.e., , otherwise we obtain a contradiction. Therefore , as required.
-
(iii)
Let satisfy
for all with on . Then
for all and so
The smallest value of satisfying this inequality is where
-
(iv)
Combining parts (ii) and (iii) gives
Therefore
as desired.
-
(v)
If , then the corresponding eigenvalue problem is
The eigenfunctions are (see Exercise Sheet 5, Q16) and the corresponding eigenvalues are , . Therefore and . In Q8 we obtained the Poincaré constant , which is obviously much bigger than the optimal constant .
-
(i)
-
13.
Uniqueness for Poisson’s equation with Robin boundary conditions. Let and be solutions of
Let . Since the PDE is linear, subtracting the equations satisfied by and gives
Multiply by and integrate by parts over :
since on . But . Therefore
The second equation implies that and hence constant (or at least constant on each connected component of ). The first equation implies that this constant must be zero. Therefore and , as required.
-
14.
Uniqueness for a more general elliptic problem. Consider the linear, second-order, elliptic PDE
(13) (i) Suppose that satisfy (13). Let . Since the PDE is linear, subtracting the equations satisfied by and gives
(14) Clearly satisfies (14). We want to show that it is the only solution. Multiply the PDE for by and integrate over :
(15) since on . Observe that
(16) by the assumption that is uniformly positive definite (take in ). Integrating by parts gives
(product rule) by the assumption that . Therefore
(17) Combining (15), (16), (17) yields
But by assumption. Therefore
and so in . Hence is constant (or at least constant on each connected component of ). But on . Therefore , as required.
(ii) The idea is the same as for (i). Let be the unique solution to the PDE with and let be the unique solution to the PDE with the matrix . Define . We need to show that in as . Taking the two PDEs, subtracting them and multiplying the resulting PDE by , we obtain
(18) Proceeding exactly as in (i), we find
Moreover, we compute (using integration by parts, since on )
So, all these arguments yield
Now, for any , Young’s inequality yields
By setting , the previous two identities imply
And by the non-negative property of , one has
We conclude by the facts that is bounded and , as .
-
15.
Uniqueness for a degenerate diffusion equation. Clearly satisfies
We use the energy method to show that it is the only positive solution. Let be any positive solution. Subtracting the PDE for from the PDE for and multiplying by gives
Now integrate over :
(Chain Rule) (Integration by parts) Therefore
But , by assumption. Hence in and so is constant in . Since on , we conclude that everywhere, as required.
-
16.
The and norms.
-
(i)
We need to check that satisfies the three properties of a norm: positivity, –homogeneity, and the triangle inequality. First we prove positivity. Let . Clearly . Suppose that and assume for contradiction that . Since is continuous, then there exists , and such that for all . Therefore
which is a contradiction. Second we check that is –homogeneous. Let . Then
as required. Finally, we prove the triangle inequality. Let . Then
Taking the square root gives the triangle inequality.
Remark: An alternative proof is to prove that the function ,
is an inner product on . It then follows that is a norm on (the norm induced by the inner product; see Definition A.16 in the lecture notes). But this is just the –norm .
Remark: The Cauchy-Schwarz inequality can be proved by considering the quadratic polynomial
Since is non-negative, then it must have non-positive discriminant, i.e., if , then . It is easy to check that this condition is exactly the Cauchy-Schwarz inequality.
-
(ii)
We will prove that the function defined by
is an inner product on . It then follows that
is a norm on (see Definition A.16 in the lecture notes). It is clear that is symmetric and bilinear and that for all . Suppose that . Then and in particular . Therefore by part (i).
-
(iii)
This is similar to part (ii). We will prove that the function defined by
is an inner product on . It is clear that is symmetric and bilinear and that for all . Suppose that . Then and in particular . Therefore by part (i) and so is a constant function. But and hence , as required.
-
(iv)
We need to find constants such that
Let . We have
Therefore . On the other hand,
where is the Poincaré constant. Therefore we can take .
-
(i)
-
17.
Continuous dependence. Let satisfy
Multiplying the PDE by and integrating over gives
by definition of the –norm. Therefore
where we have used the Cauchy-Schwarz inequality and the fact that for all . Cancelling one power of from both sides gives the desired result:
with .
Remark: Note that this estimate degenerates as tends to ( as ). If or is small then a better estimate can be obtained using the Poincaré inequality: As above
Therefore
where is the Poincaré constant. Cancelling one power of from both sides gives
with .
-
18.
Continuous dependence with a first-order term.
-
(a)
Let satisfy
(19) Multiply the PDE by and integrate over :
by the Cauchy-Schwarz inequality.
-
(b)
Let . Then
by the Young inequality.
-
(c)
Combining parts (a) and (b) gives
- (d)
- (e)
-
(a)
-
19.
Neumann boundary conditions for variational problems.
-
(i)
Let be a minimiser of . For any , , define Then since the sum of functions is . Let . Note that when . Therefore is minimised by since is minimised by . Hence
Therefore
(21) as required.
-
(ii)
First choose a test function such that on . Since , we can integrate by parts in (21) to obtain
because on . Since this holds for all test functions such that on , the Fundamental Lemma of the Calculus of Variations implies that
(22) as required. We still need to show that satisfies the Neumann boundary condition. Now take any test function in (21) and integrate by parts as before to obtain
Since this holds for all , then on , as required.
-
(i)
-
20.
The –Laplacian operator.
-
(i)
Let minimise . For any , , define Observe that vanishes on the boundary of since both and vanish there. Also since the sum of functions is . Hence . Define . Now when . Therefore is minimised by since is minimised by . We have reduced the problem of minimising the functional to minimising the function of one variable . Since is minimised at ,
(23) where the differentiation was performed using the Chain Rule and the fact that
Recall the integration by parts formula
By applying this with , , we can rewrite equation (23) as
But on since . Therefore
Since is arbitrary, the Fundamental Lemma of the Calculus of Variations gives
Therefore
as required. Note that on by definition of .
-
(ii)
Multiply the PDE by and integrate by parts over to obtain
(24)
-
(i)
-
21.
The minimal surface equation: PDEs and soap films. Let be a minimiser of . Let and with on . Define . Then since the sum of continuously differential functions is continuously differentiable and, if , then
as required. Define by . Then and so is a minimum point of since is a minimum point of . Therefore
This means that is a weak solution of the minimal surface equation. Since , then we can integrate by parts to obtain
since on . This holds for all with . Therefore satisfies the minimal surface equation
by the Fundamental Lemma of the Calculus of Variations (Lemma 3.20).
-
22.
Homogenization and the calculus of variations.
-
(i)
Let minimise . For any and any such that , define Then
and similarly . Therefore . Define . Now when . Therefore the minimum of is attained at since the minimum of is attained at . We have reduced the problem of minimising the functional to minimising the function of one variable . Since is minimised at ,
(25) Since , we can use integration by parts to rewrite equation (25) as
But this holds for all such that . Therefore by the Fundamental Lemma of the Calculus of Variations
as required. Note that satisfies the Dirichlet boundary conditions by definition of .
-
(ii)
Recall from Q2(ii) that if is –periodic, then for any interval ,
(26) Applying (26) with , , , on , gives the desired result:
-
(iii)
Observe that is just the one-dimensional Dirichlet energy with an additional constant in the first term. It follows from Dirichlet’s Principle (see the lecture notes) that satisfies the Poisson equation
In Q2 we showed that , where satisfies
where
Since in general, it follows that and hence
as required.
In fact it can be shown that as follows:
where we have used the Cauchy-Schwarz inequality. It follows that the –limit is less than or equal to the pointwise limit .
-
(i)