Partial Differential Equations III/IV
[6pt] Exercise Sheet 4
-
1.
Green’s functions. Consider Poisson’s equation in one dimension with mixed Dirichlet-Neumann boundary conditions:
where . Find the unique solution of this equation. Write your solution in the form
-
2.
Homogenization. Let be constant and be –periodic, i.e., for all . For , define the –periodic function . Consider the steady diffusion equation
(1) (2) where is continuous. This models the equilibrium temperature distribution in a metal bar with heat source , where the bar is made of a nonhomogeneous material. The thermal conductivity depends on position and represents a metal bar composed of repeating segments of length . We will discover that, for small , the bar behaves as if it were made of a homogenous material with constant thermal conductivity . You might guess that is some sort of average of , but what is the correct notion of average? This is called a homogenization problem. Homogenization is an active research area.
- (i)
-
(ii)
Let . Prove that , where
and where
is the average of over one period.
Hint: Use the following deep result, which you do not need to prove: Let be –periodic. For any interval ,
(3) where is the average of over one period, i.e., . This also holds under the weaker assumption that . Define . We say that converges weak- in to as , but that’s another story for a functional analysis course.
-
(iii)
Write in the form
for some symmetric Green’s function , which you should determine.
-
(iv)
Use part (ii) to show that satisfies the steady diffusion equation
where the thermal conductivity is the constant
-
(v)
Observe that is the reciprocal of the average of the reciprocal of . In general this is not the same as the average of , as we now illustrate. Let
and extend by periodicity to the real line. This represents a composite metal bar composed of segments of two homogeneous materials with different thermal conductivities. Compute and . Verify that .
-
(vi)
Bonus, optional question (hard): Prove (3). This is a form of the Riemann-Lebesgue Lemma. It is a generalisation of the Riemann-Lebesgue Lemma for Fourier series and the Fourier transform.
Hint: Start by writing
-
3.
Radial symmetry of Laplace’s equation on . Let
be the set of real, –by– orthogonal matrices, which represent rotations and reflections of . Let be a harmonic function and let . Define by . Prove that is also harmonic.
-
4.
Fundamental solution of Poisson’s equation in 3D. Let be the fundamental solution of Poisson’s equation in :
-
(i)
Let . Compute .
Hint: Use spherical polar coordinates or, simpler, the following formula:
-
(ii)
Prove that .
-
(iii)
Prove that .
-
(iv)
Prove that .
-
(i)
-
5.
Fundamental solution of Poisson’s equation in 1D. Let be twice continuously differentiable with compact support. Define
We call the fundamental solution of Poisson’s equation in . Define . Prove that satisfies
Hint: Write
Now integrate by parts. Unlike for the case of Poisson’s equation in , , you do not need to remove a ball of radius around the origin since does not have a singularity at the origin in 1D. In fact, unlike in higher dimensions, is continuous in 1D.
-
6.
The function spaces and . Let , . Let . Find all the values of for which
-
(i)
,
-
(ii)
,
-
(iii)
,
-
(iv)
.
-
(i)
-
7.
Properties of the convolution. Let , . Prove
-
(i)
for all ;
-
(ii)
if , then ;
-
(iii)
the convolution is commutative:
Remark: It can be shown that if . Consequently is an algebra.
-
(i)
-
8.
The Poincaré inequality for functions that vanish on the boundary. Prove that there exists a constant such that
for all satisfying .
Hint: The proof is similar to, and simpler than, the version we proved in Section 4.3. -
9.
The Poincaré inequality on unbounded domains.
-
(i)
Construct a sequence such that and
This means that the Poincaré inequality does not hold on , i.e, there does not exists any such that
for all with .
-
(ii)
Let be the unbounded domain . Prove that there exists such that
for all with and with for all . More generally, the Poincaré inequality is true if is bounded between two parallel hyperplanes (lines in 2D, planes in 3D, etc). In this example is bounded between the lines and .
-
(i)
-
10.
The Poincaré constant depends on the domain. Let satisfy
(4) for all with . By using a change of variables, use (4) to prove that
(5) for all with , where
Remark: Those with a good physical intuition will see that must have units of length squared, otherwise the units in equation (5) do not match: if is dimensionless, then has units of length whereas has units of length.
-
11.
Eigenvalues of : Can you hear the shape of a drum? Let be open and bounded with smooth boundary. Let be a smooth eigenfunction of that vanishes on , which means and
for some . We say that is the eigenvalue associated to the eigenfunction . Use the energy method to prove that is real and that .
Hint: Start by multiplying the PDE by , the complex conjugate of .
Remark: This eigenvalue problem arises if you seek a solution of the form of the wave equation with clamped boundary conditions, which models small vibrations of a drum of shape . The eigenvalues are related to the principal frequencies of the drum by . It can be shown that there are countably-many eigenvalues . Moreover, H. Weyl showed that the eigenvalues determine the area of ; you can hear the area of a drum. In 1966 M. Kac asked whether the eigenvalues determine the shape of ; can you hear the shape of a drum? This was disproved in 1992 by Gordon, Webb and Wolpert, who constructed two distinct, non-convex polygons with the same principal frequencies. As a final twist, it is possible to hear the shape of a convex drum; two distinct convex sets have different principal frequencies.
-
12.
The optimal Poincaré constant and eigenvalues of .
-
(i)
Use the energy method to show that if the pair , , satisfies
(6) then
(7) This is the Rayleigh quotient formula for the eigenvalue in terms of the eigenfunction . It can be shown that there are countably many eigenvalues and that .
Remark: The Rayleigh quotient formula for the matrix eigenvalue problem isThis has the same form of (7) but with the –inner product in (7) replaced by the dot product.
-
(ii)
Define
and define the functional by
Suppose that minimises , i.e.,
Prove that
where is the smallest eigenvalue of (6), and that
Remark: We have shown that the minimum value of the Rayleigh quotient is the smallest eigenvalue of the operator on , and that is minimised by the corresponding eigenfunction. This is analogous to the result that if is a symmetric positive definite matrix, then the minimum value of the Rayleigh quotient is the smallest eigenvalue of , and it is minimised by the corresponding eigenvector. (Recall also that the maximum value of the Rayleigh quotient is the largest eigenvalue of ). Equivalently, the minimum value of the quadratic form over the sphere is the smallest eigenvalue of .
-
(iii)
The optimal Poincaré constant is the smallest value of such that
(8) for all with on . Let us denote this value of by . Show that
-
(iv)
Combine parts (ii) and (iii) to conclude that the optimal Poincaré constant is
-
(v)
Use part (iv) to show that the optimal Poincaré constant for the domain is . Compare this with the constant you obtained in Q8 with , .
Remark: The optimal constant can also be obtained using Fourier series.
-
(i)
-
13.
Uniqueness for Poisson’s equation with Robin boundary conditions. Let be open and bounded with smooth boundary. Let . Use the energy method to show that there is at most one smooth solution of
where denotes the outward-pointing unit normal vector field to . This type of boundary condition is called a Robin boundary condition.
-
14.
Uniqueness and stability for a more general elliptic problem. Consider the linear, second-order, elliptic PDE
(9) where is open and bounded with smooth boundary, , , and . Assume that is nonnegative, div, and is uniformly positive definite, i.e., there exists a constant such that for all , .
-
(i)
Prove that (9) has at most one solution .
-
(ii)
Let be a sequence of matrix valued functions satisfying the previous conditions (in particular they are all uniformly positive definite with the same ). The rest of the data is fixed and satisfy all the previous assumptions. Suppose that is the unique solutions to
Show that if uniformly in as and if the problem with the limit matrix has a unique solution , then in , as .
Remarks: Uniqueness may fail if is negative; see the PDEs exam from May 2017, Q4(b). Another obstacle to uniqueness is unbounded domains; see Exercise Sheet 5. We say that the PDE (9) has divergence form, which is the most convenient form for energy methods (compare (9) with the general form of elliptic PDEs given in Definition 4.1).
-
(i)
-
15.
Uniqueness for a degenerate diffusion equation. Let be a constant. Show that the following steady degenerate diffusion equation has a unique positive solution:
Remark: Observe that with . We call the equation the degenerate diffusion equation since the diffusion coefficient vanishes when .
-
16.
The and norms. Let
-
(i)
Prove that is a norm on .
Hint: The only difficulty is proving the triangle inequality. Write
and use the Cauchy-Schwarz inequality.
-
(ii)
Prove that is a norm on .
-
(iii)
Prove that is a norm on . Is it a norm on ?
-
(iv)
Prove that the norms and are equivalent on , which means that there exist constants such that
Hint: Use the Poincaré inequality to find .
Remark: If two norms are equivalent, then a sequence converges in one norm if and only if it converges in the other.
-
(i)
-
17.
Continuous dependence. Let be open and bounded with smooth boundary. Let satisfy
where , is uniformly elliptic (see Q9), and is a constant. Prove that there exists a constant such that
-
18.
Continuous dependence with a first-order term. Let be open and bounded with smooth boundary. Let , be constants and let , be continuous. Let satisfy
(10) -
(a)
Prove that
-
(b)
Prove that for all
Hint: You may use the Young inequality, which states that
-
(c)
Prove that for all
-
(d)
Find a constant such that if , then
for some constant .
-
(e)
Show that if , then is the only solution of (10).
-
(a)
-
19.
Neumann boundary conditions for variational problems. Let be open and bounded with smooth boundary. Define by
Suppose that minimises :
-
(i)
Show that
This is the weak formulation of Poisson’s equation with zero Neumann boundary conditions, as the following part demonstrates:
-
(ii)
Show that, if in addition , then
where denotes the outward-pointing unit normal vector field to .
Hint: First choose test functions such that on . Use this to establish that . Then choose any test function and show that the boundary condition holds.
Remark: Observe that the Neumann boundary condition arises naturally without including it in the domain of (cf. the case of Dirichlet boundary conditions in Section 4.5, where the boundary condition is included in the domain of the energy functional). Consequently Neumann boundary conditions are sometimes referred to as natural boundary conditions.
-
(i)
-
20.
The –Laplacian operator. Let be open and bounded with smooth boundary. Let
For , define by
We met the case in Section 4.5 and the previous question. Suppose that minimises :
-
(i)
Prove that satisfies the PDE
where is the –Laplacian operator, which is defined by By taking we recover the regular Laplacian operator: .
-
(ii)
Show that
-
(i)
-
21.
The minimal surface equation: PDEs and soap films. This question is adapted from the PDEs exam, May 2017. Let be open and bounded with smooth boundary. Let be a given smooth function and let
Define by
Observe that is the area of the surface , i.e., is the surface area of the graph of . Suppose that the graph of has minimal surface area amongst all graphs with given boundary :
Show that satisfies the minimal surface equation
-
22.
Homogenization and the calculus of variations. In this question we revisit the homogenization problem from Q2 from the viewpoint of the calculus of variations. Let be constants. Define
Let be constant and be continuously differentiable. Let be continuous. Define the energy functional by
Observe that we recover the Dirichlet energy when .
-
(i)
Suppose that minimises :
Show that satisfies
-
(ii)
Now assume that is –periodic and consider the energy
Show that
where
This means the nonhomogeneous energy converges pointwise to the homogeneous energy .
-
(iii)
Let . By part (i), if is the minimiser of , then satisfies (1), (2) for . Let be the minimiser of . Show that
Interpretation: We have shown that converges pointwise to , but the minimiser of does not converge to the minimiser of . The moral of the story is that pointwise convergence is not the ‘correct’ notion of convergence when considering energy functionals. The correct notion is something called –convergence. It can be shown that –converges in a suitable sense to
where was defined in Q2(iv). It is easy to check that the minimiser of converges to the minimiser of . The subject of –convergence goes beyond the scope of this course, but the important property of –convergence is that minimisers converge to minimisers. This means that if you are modelling a system with an energy functional and you want to simplify the functional by sending a large parameter or a small parameter , then you should compute the –limit, not the pointwise limit.
-
(i)