Skip to the Main Content

Note:These pages make extensive use of the latest XHTML and CSS Standards. They ought to look great in any standards-compliant modern browser. Unfortunately, they will probably look horrible in older browsers, like Netscape 4.x and IE 4.x. Moreover, many posts use MathML, which is, currently only supported in Mozilla. My best suggestion (and you will thank me when surfing an ever-increasing number of sites on the web which have been crafted to use the new standards) is to upgrade to the latest version of your browser. If that's not possible, consider moving to the Standards-compliant and open-source Mozilla browser.

March 25, 2018

On the Magnitude Function of Domains in Euclidean Space, II

Posted by Simon Willerton

joint post with Heiko Gimperlein and Magnus Goffeng.

In the previous post, On the Magnitude Function of Domains in Euclidean Space, I, Heiko and Magnus explained the main theorem in their paper

(Remember that here a domain XX in R nR^n means a subset equal to the closure of its interior.)

The main theorem involves the asymptoic behaviour of the magnitude function X(R)\mathcal{M}_X(R) as RR\to\infty and also the continuation of the magnitude function to a meromorphic function on the complex numbers.

In this post we have tried to tease out some of the analytical ideas that Heiko and Magnus use in the proof of their main theorem.

Heiko and Magnus build on the work of Mark Meckes, Juan Antonio Barceló and Tony Carbery and give a recipe of calculating the magnitude function of a compact domain X nX\subset \mathbb{R}^n (for n=2m1n=2m-1 an odd integer) by finding a solution to a differential equation subject to boundary conditions which involve certain derivatives of the function at the boundary X\partial X and then integrating over the boundary certain other derivatives of the solution.

In this context, switching from one set of derivatives at the boundary to another set of derivatives involves what analysts call a Dirichlet to Neumann operator. In order to understand the magnitude function it turns out that it suffices to consider this Dirichlet to Neumann operator (which is actually parametrized by the scale factor in the magnitude function). Heavy machinary of semiclassical analysis can then be employed to prove properties of this parameter-dependent operator and hence of the magntiude function.

We hope that some of this is explained below!

[Remember that throughout this post we have n=2m1n=2m-1 is an odd positive integer.]

The work of Meckes and Barceló-Carbery

As a reader of this blog you might well know that magnitude of finite metric spaces is usually defined using weightings. Mark Meckes showed that the natural extension of magnitude to infinite subsets of Euclidean space can be defined using potential functions.

Before going anywhere, however, recall that the Laplacian operator Δ\Delta is the differential operator on functions on n\mathbb{R}^n given by Δf= i=1 n 2fx i 2\Delta f=\sum_{i=1}^n \frac{\partial ^2f}{\partial x_i^2}.

Now, for XX a compact subset of n\mathbb{R}^n with smooth boundary, a potential function hh for XX is a function h: nh\colon \mathbb{R}^n\to \mathbb{R} with properties including the following:

  • h=1h= 1 on XX;
  • (idΔ) mh=0(\id- \Delta)^m h = 0 (weakly) on nX\mathbb{R}^n\setminus X;
  • hh is m1m-1 times differentiable on n\mathbb{R}^n, and the mmth derivative exists in an L 2L^2-sense;
  • h(x)0h(x)\to 0 as xx\to \infty.

You can see an example below in the next section.

Barceló and Carbery built on the results of Meckes to show that for a compact convex domain XX in n\mathbb{R}^n with smooth boundary the following recipe can be used to calculate the magnitude of XX.

First define 𝒟 i\mathcal{D}^i to be the order ii differential operator on the boundary by

𝒟 2j=Δ j,𝒟 2j+1=νΔ j, \mathcal{D}^{2j}= \Delta^{j},\,\,\,\,\, \mathcal{D}^{2j+1}=\textstyle\frac{\partial}{\partial \nu}\Delta^{j},

where ν\textstyle\frac{\partial}{\partial \nu} means the derivative in the normal direction to the boundary.

The Barceló-Carbery Recipe for Magnitude. Suppose X nX\in \mathbb{R}^n is a compact domain with smooth boundary.

  1. Find a solution u: nXu\colon \mathbb{R}^n\setminus X\to \mathbb{R} with u(x)0u(x)\to 0 as xx\to \infty of the differential equation (IdΔ) mu=0on nX (Id-\Delta)^m u=0\,\,\,\, \text{on}\,\,\mathbb{R}^n\setminus X subject to the boundary conditions u=1,𝒟 1(u)=0,𝒟 2(u)=0,,𝒟 m1(u)=0,onX. u=1,\, \mathcal{D}^1(u) =0,\, \mathcal{D}^2(u) =0,\, \dots, \mathcal{D}^{m-1}(u) =0, \,\,\,\, \text{on}\,\, \partial X.

  2. The magnitude is then calculated by mag(X)=1n!ω n(vol(X)+ m/2<jm(1) j(mj) X𝒟 2j1(u)ds). \text{mag}(X)=\frac{1}{n!\,\omega_n}\left(\text{vol}(X)+\sum_{m/2\lt j\le m}(-1)^j\binom{m}{j}\int_{\partial X} \mathcal{D}^{2j-1}(u)\,\mathrm{d}{s}\right).

Barcelo and Carbery actually stated their result for convex domains, but if we assume smoothness of the boundary then we can drop the convexity assumption.

The potential function hh of XX is related to the uu in the recipe by extending it to all of n\mathbb{R}^n by taking it to be 11 on XX.

Let’s have a look at a simple example that we’ll return to through this post.

A one-dimensional example

Consider the union of two intervals in the line: X:=[a 1,a 2][a 3,a 4]X:=[a_1, a_2]\cup[a_3, a_4]\subset \mathbb{R} for a 1<a 2<a 3<a 4a_1\lt a_2\lt a_3 \lt a_4. The differential equation to solve in the Barceló-Carbery recipe is then

uu=0on X=(,a 1][a 2,a 3][a 4,), u-u''=0 \,\,\,\, \text{on }\,\,\mathbb{R}\setminus X=(-\infty, a_1]\cup [a_2,a_3]\cup [a_4,\infty),

and the boundary conditions are

u(x)=1for x{a 1,a 2,a 3,a 4}. u(x)=1\,\,\,\,\text{for }\,\,x\in \{a_1, a_2, a_3, a_4\}.

This is easy to solve by hand and you find the solution

u(x)={e (a 1x) x(,a 1], (e (xa 2)+e (a 3x))/(e (a 3a 2)+1) x[a 2,a 3], e (xa 4) x[a 4,). u(x)=\begin{cases} e^{-(a_1-x)} & x\in (-\infty, a_1], \\ (e^{-(x-a_2)}+e^{-(a_3-x)})/(e^{-(a_3-a_2)}+1) & x\in [a_2, a_3], \\ e^{-(x-a_4)} & x\in [a_4, \infty). \end{cases}

Here is the graph of the potential function.

The graph of the potential function

Now according to Barceló and Carbery’s recipe we can calculate the magnitude as

mag(X) =12(vol(X)(u(a 1)+u(a 2)u(a 3)+u(a 4))) \begin{aligned} \mathrm{mag}(X) & =\frac{1}{2}\left(\mathrm{vol}(X)-(-u'(a_1)+ u'(a_2)-u'(a_3)+u'(a_4))\right) \end{aligned}

But it is easy to compute from the formula for uu above that

u(a 1)=1=u(a 4),u(a 2)=tanh(a 3a 22)=u(a 3) u'(a_1)=1=-u'(a_4), \quad -u'(a_2)=\tanh\left(\frac{a_3-a_2}{2}\right)=u'(a_3)

and so

mag(X) =12(a 2a 1+a 4a 3)+1+tanh(a 3a 22) \begin{aligned} \mathrm{mag}(X)&=\tfrac{1}{2}(a_2-a_1+a_4-a_3)+1 +\tanh(\frac{a_3-a_2}{2}) \end{aligned}

which we can write as

mag(X) =12vol(X)+χ(X)2exp(a 3a 2)+1. \begin{aligned} \mathrm{mag}(X) &= \tfrac{1}{2}\mathrm{vol}(X)+\chi(X)-\frac{2}{\exp(a_3-a_2)+1}. \end{aligned}

A key point to note here is that to calculate the magnitude we don’t actually need to know the whole potential function uu, we only need to know certain of its derivatives at the boundary. So we start with a differential equation, specify sufficiently many derivatives at the boundary to give a unique solution and then find values of other derivatives at the boundary. This is a process which is well studied in the area of boundary value problems and is embedded in the notion of the Dirichlet to Neummann operator which we now look at.

The Dirichlet to Neumann operator

As you surely know, when solving a differential equation, you impose boundary conditions in order to pin down the solution. You might impose different boundary conditions in different situations. For instance, the classical Dirichlet boundary conditions for a problem of second order fix the value of the function on the boundary, whereas the classical Neumann boundary conditions fix the normal derivative of the function on the boundary.

For the calculating magnitude, the boundary value problem is of order 2m2m, not of order 22. We think of the boundary conditions f=1,𝒟 1(f)=0,𝒟 2(f)=0,,𝒟 m1(f)=0f=1,\, \mathcal{D}^1(f) =0, \, \mathcal{D}^2(f) =0, \,\dots, \mathcal{D}^{m-1}(f) =0, which involves derivatives of order 00 up to m1m-1, as analogues of Dirichlet boundary conditions. To compute the magnitude we need to determine the derivatives of the solution of order mm up to 2m12m-1, which we think of as analogues of Neumann boundary conditions.

Given Dirichlet boundary conditions we want to determine the corresponding Neumann boundary conditions. Let’s think what this means.

If you have a differential equation Lf=0L f =0 on a domain XX then specifying the boundary condition means imposing a set of equations of the form

(δ 1f)(x)=v 1(x),,(δ pf)(x)=v p(x),for all xX (\delta_1f)(x)=v_1(x),\,\,\dots,\,\,(\delta_p f)(x)=v_p(x), \,\, \text{for all }\,\,x\in \partial X

where each δ i\delta_i is a differential operator on the boundary X\partial X and each v iFun(X)v_i\in \mathrm{Fun}(\partial X) belongs to a suitable space of functions on the boundary. (We will avoid technicalities and complicated notation by using Fun(X)Fun(\partial X) to stand for some space of functions, which might vary depending on context.)

When the boundary conditions give a unique solution to the differential equation – such as in the Barceló-Carbery Recipe – then for any other set {δ˜ j} j=1 q\{\tilde {\delta}_j\}_{j=1}^q of differential operators on the boundary there is a map, the Dirichlet to Neumann operator, between tuples of function on the boundary:

Λ: i=1 pFun(X) j=1 qFun(X); i=1 pv i j=1 qδ˜ ju v, \begin{aligned} \Lambda\colon \bigoplus_{i=1}^p \mathrm{Fun}(\partial X) &\to \bigoplus_{j=1}^q \mathrm{Fun}(\partial X); \\ \bigoplus_{i=1}^p v_i &\mapsto\bigoplus_{j=1}^q {\tilde\delta}_j u_{\mathbf{v}}, \end{aligned}

where u vu_{\mathbf{v}} is the unique solution to Lu=0L u=0 subject to the boundary conditions δ iu=v i\delta_i u=v_i, i=1,,pi=1,\ldots, p.

This Dirichlet to Neumann operator will be a key ingredient in our approach to the parameter-dependent boundary problem for the magnitude function below.

In our toy one-dimensional example we have a single differential operator δ 1=id\delta_1= \id and v 11v_1\equiv 1, so this is a classical Dirichlet boundary condition; and we have δ˜ 1\tilde\delta_1 being the normal derivative to the boundary and therefore this is a classical Neumann boundary condition.

Let’s see what this operator Λ\Lambda is in this example.

The Dirichlet to Neumann operator in our toy example

The boundary of XX consists of the four points {a 1,a 2,a 3,a 4}\{a_1, a_2, a_3, a_4\}, and we can identify the space of functions on the boundary Fun(X)\mathrm{Fun}(\partial X) with 4\mathbb{C}^4. We define the linear map Λ:Fun(X)Fun(X)\Lambda\colon\mathrm{Fun}(\partial X)\to \mathrm{Fun}(\partial X), ie. Λ: 4 4\Lambda\colon\mathbb{C}^4\to \mathbb{C}^4 as

Λ((z 1 z 2 z 3 z 4))(u(a 1) u(a 2) u(a 3) u(a 4)), \Lambda\left(\begin{pmatrix} z_1\\ z_2\\ z_3\\ z_4\end{pmatrix} \right) \coloneqq \begin{pmatrix} u'(a_1)\\ -u'(a_2)\\ u'(a_3)\\ -u'(a_4) \end{pmatrix} \, ,

where the function uu solves the boundary value problem

u=uinX and(u(a 1) u(a 2) u(a 3) u(a 4))=(z 1 z 2 z 3 z 4). {u'}'=u\,\,\,\,\text{in}\,\, \mathbb{R}\setminus X\,\,\text{ and}\,\,\,\, \begin{pmatrix} u(a_1)\\ u(a_2)\\ u(a_3)\\ u(a_4)\end{pmatrix} = \begin{pmatrix} z_1\\ z_2\\ z_3\\ z_4\end{pmatrix}\, .

It is not difficult to compute that

Λ=(1 0 0 0 0 coth(a 3a 2) csch(a 3a 2) 0 0 csch(a 3a 2) coth(a 3a 2) 0 0 0 0 1). \Lambda= \begin{pmatrix} 1&0&0&0\\ 0& \coth(a_3-a_2)&-\mathrm{csch}(a_3-a_2)&0\\ 0& -\mathrm{csch}(a_3-a_2)&\coth(a_3-a_2)&0\\ 0&0&0& 1 \end{pmatrix}\, .

Using the Barceló-Carbery recipe we have that the magnitude can be obtained from the sum of the entries of the matrix of Λ\Lambda: writing 1=(1,1,1,1) T\vec{1}=(1,1,1,1)^T,

mag(X)=vol(X)2+121,Λ1 4=vol(X)2+1+tanh(a 3a 22), \mathrm{mag}(X)=\frac{\mathrm{vol}(X)}{2}+\frac{1}{2}\langle \vec{1},\Lambda\vec{1}\rangle_{\mathbb{C}^4}=\frac{\mathrm{vol}(X)}{2}+1+\tanh\left(\frac{a_3-a_2}{2}\right)\, ,

as we had before.

Of course, in this case we did calculate the potential function in order to calculate Λ\Lambda. For domains in higher dimensions it is rarely possible to compute the potential function. This is the reason to introduce heavier guns from global analysis allowing us to study the operator Λ\Lambda without explicitly solving the problem.

What we are really interested is the magnitude function, its meromorphicity and asymptotic behaviour, so we need to study the above boundary value problems with a parameter which will represent the scale factor.

Introducing a parameter

Remember that the magnitude function, X\mathcal{M}_X, is defined in terms of the magnitude of the dilates of XX, ie. X(R)mag(RX)\mathcal{M}_X(R)\coloneqq\mathrm{mag}(R\cdot X) for R>0R\gt 0, where RXR\cdot X is the same space XX but with the metric scaled up by a factor of RR.

The Barceló-Carbery recipe for the magnitude from above can be generalized to include the scale factor RR and in such a way so that it is on an equal footing with the derivatives, essentially by replacing (IdΔ)(Id-\Delta) with (R 2Δ)(R^2-\Delta). This approach is well studied in the literature on parameter-dependent pseudo-differential operators.

First define 𝔻 R i\mathbb{D}_R^i to be the order ii differential operator on the boundary X\partial X given by 𝔻 R 2j=(R 2Δ) j,𝔻 R 2j+1=ν(R 2Δ) j. \mathbb{D}_R^{2j}= (R^2-\Delta)^{j},\,\,\,\, \mathbb{D}_R^{2j+1} = \textstyle\frac{\partial}{\partial \nu}(R^2-\Delta)^{j}.

The Gimperlein-Goffeng Recipe for the Magnitude Function. Suppose that X nX\subset \mathbb{R}^n is a compact domain with smooth boundary.

  1. Find a solution u R: nXu_R\colon \mathbb{R}^n\setminus X\to \mathbb{R} with u R(x)0u_R(x)\to 0 as xx\to \infty of the differential equation (R 2Δ) mu=0on nX (R^2-\Delta)^m u=0\,\,\,\,\text{on }\,\,\mathbb{R}^n\setminus X subject to the boundary conditions on 𝔻 R 0(u),,𝔻 R m1(u)\mathbb{D}_R^0(u),\dots, \mathbb{D}_R^{m-1}(u): 𝔻 R 2i(u)=R 2i,𝔻 R 2i+1(u)=0,on X. \mathbb{D}_R^{2i}(u) =R^{2 i},\,\, \mathbb{D}_R^{2i+1}(u) =0, \,\,\,\, \text{on } \partial X.

  2. The magnitude is then calculated by mag(RX)=1n!ω n(vol(X)R n m/2<jmR n2j X𝔻 R 2j1u RdS). \text{mag}(R\cdot X)=\frac{1}{n!\,\omega_n}\left(\text{vol}(X)R^n-\sum_{m/2\lt j\le m} R^{n-2j}\int_{\partial X} \mathbb{D}_R^{2j-1}u_R \,\mathrm{d}{S}\right).

The eagle-eyed amongst you will notice that setting R=1R=1 does not immediately recover the Barceló-Carbery recipe. However, you can recover that with some algebraic manipulation and binomial identities.

Again we can use the Dirichlet to Neumann operator, but note that this will depend on a parameter RR. We think of Λ(R)\Lambda(R) as an operator valued function of the scaling parameter RR. If we start with the values of the differential operators 𝔻 R 0(u),,𝔻 R m1(u)\mathbb{D}_R^0(u),\dots, \mathbb{D}_R^{m-1}(u) on the boundary it should return the values of the operators 𝔻 R m(u),𝔻 R m+2(u),,𝔻 R n(u)\mathbb{D}_R^{m'}(u),\,\,\mathbb{D}_R^{m'+2}(u),\,\,\dots, \mathbb{D}_R^{n}(u), where m=mm'=m if mm is odd and m=m+1m'=m+1 is mm is even. By the formula above we can use this operator to calculate the magnitude function.

Let’s look at the case of our toy example again.

The parameter-dependent operator in the toy example

In our running example of two disjoint intervals on the real line, X:=[a 1,a 2][a 3,a 4]X:=[a_1, a_2]\cup[a_3, a_4]\subset \mathbb{R}, you can calculate to find

Λ(R)=R(1 0 0 0 0 coth(R(a 2b 1)) csch(R(a 2b 1)) 0 0 csch(R(a 2b 1)) coth(R(a 2b 1)) 0 0 0 0 1). \Lambda(R)= R\begin{pmatrix} 1&0&0&0\\ 0& \coth(R(a_2-b_1))&-\mathrm{csch}(R(a_2-b_1))&0\\ 0& -\mathrm{csch}(R(a_2-b_1))&\coth(R(a_2-b_1))&0\\ 0&0&0& 1 \end{pmatrix}\,.

Again, writing 1=(1,1,1,1) T\vec{1}=(1,1,1,1)^T, we compute the magnitude, using the Gimperlein-Goffeng recipe, as the sum of all the entries:

X(R)=vol(X)2R+12R1,Λ(R)1 4=vol(X)2R+1+tanh(R(a 2b 1)2). \mathcal{M}_X(R)=\frac{\mathrm{vol}(X)}{2}R+\frac{1}{2R}\langle \vec{1},\Lambda(R)\vec{1}\rangle_{\mathbb{C}^4}=\frac{\mathrm{vol}(X)}{2}R+1+\tanh\left(\frac{R(a_2-b_1)}{2}\right).

It is worth noting that you can see that the operator Λ(R)\Lambda(R) depends meromorphically on RR\in \mathbb{C}, rather than just being defined for R>0R\gt 0, and Λ(R)\Lambda(R) an asymptotic expansion as Re(R)\mathrm{Re}(R)\to \infty. Therefore, the same holds for X(R)\mathcal{M}_X(R).

Proving the main theorem!

As described in the previous post, the main theorem of the paper is about a meromorphic extension of the magnitude function and about the asymptotic behaviour of the magnitude function X(R)\mathcal{M}_X(R) as RR\to \infty. As we’ve seen above, the magnitude function can be calculated from the parameter-dependent Dirichlet to Neumann operator Λ(R)\Lambda(R). Now heavy machinery from geometric and semiclassical analysis – such as meromorphic Fredholm theorem and parameter-dependent pseudo-differential operators – can be used to extend Λ(R)\Lambda(R) to a meromorphic operator valued function and study its asymptotic expansion as RR \to \infty. The properties of the magnitude function then follow.

That is probably enough for now, but in the next post, there should be a slightly less trivial example and some thoughts and comments of a more general nature.

Posted at March 25, 2018 1:46 PM UTC

TrackBack URL for this Entry:   https://golem.ph.utexas.edu/cgi-bin/MT-3.0/dxy-tb.fcgi/3029

5 Comments & 0 Trackbacks

Re: On the Magnitude Function of Domains in Euclidean Space, II

Thanks for the post!

Would someone like to explain — very roughly! — what semiclassical analysis is? It’s a funny name: you don’t hear people talking about semiclassical music. But it seems to have a particular meaning.

Posted by: Tom Leinster on March 27, 2018 12:57 AM | Permalink | Reply to this

Re: On the Magnitude Function of Domains in Euclidean Space, II

Hopefully one of the authors of the post will come along and give a more complete answer, but the context in which I’m familiar with the term semiclassical is in physics, where it usually refers to doing things in a not-quite-quantum way. One technique that goes by this name is to do calculations that are hard to do for positive \hbar in the limit as 0\hbar \to 0.

After Heiko and Magnus wrote their paper, I asked a geometric analyst friend of mine what “semiclassical analysis” means to him. I didn’t retain the details, but I got enough to understand that it does indeed refer to things that are analogous to — and named in reference to — that technique in physics. Somehow the 0\hbar \to 0 limit there relates to the RR \to \infty limit in the differential operators (\hbar shows up in quantum mechanics as a coefficient in operators in the Schrödinger equation).

Take everything I just said (and anything else you ever hear me say about geometric analysis and/or physics) with a large grain of salt. But that’s what I think I heard someone say to me.

Posted by: Mark Meckes on March 27, 2018 3:30 AM | Permalink | Reply to this

Re: On the Magnitude Function of Domains in Euclidean Space, II

I see. Thanks. So in physics, classical would be =0\hbar = 0, quantum would be >0\hbar \gt 0, and semiclassical would be 0\hbar \to 0. That makes “semiclassical” sound something like “infinitesimal” in this context.

Posted by: Tom Leinster on March 28, 2018 3:03 AM | Permalink | Reply to this

Re: On the Magnitude Function of Domains in Euclidean Space, II

Dear Tom,

Apologies for the delayed reply.

Mark’s comment already gives some hint that semiclassical analysis concerns the analysis of partial differential equations with a small, positive parameter \hbar. Here, (R 2Δ) m=R 2m(1 2Δ) m(R^2-\Delta)^m=R^{2m}(1-\hbar^2\Delta)^m with =1/R\hbar=1/R. In physics, 2Δ-\hbar^2\Delta is the quantization of kinetic energy.

For physics, semiclassical analysis provides rigorous tools to relate classical and quantum mechanics (geometric optics/stationary phase approximation), based on harmonic analysis. In mathematics, it allows the detailed study of (often linear) partial differential equations with a small parameter. The maybe simplest example is the stationary phase approximation of oscillatory integrals. Particular successes of semiclassical analysis include the heat kernel proofs of the Atiyah-Singer index theorem, as well as the mathematical analysis of problems in quantum mechanics and quantum field theory.

The introduction to Maciej Zworski’s book provides some further context and an overview over current topics of interest.

Posted by: Heiko Gimperlein on March 29, 2018 8:37 PM | Permalink | Reply to this

Re: On the Magnitude Function of Domains in Euclidean Space, II

A very beautiful analytic formula for the magnitude!

Posted by: Bruce Bartlett on March 27, 2018 10:17 PM | Permalink | Reply to this

Post a New Comment