Conformal radius

From testwiki
Jump to navigation Jump to search

Template:More citations needed

In mathematics, the conformal radius is a way to measure the size of a simply connected planar domain D viewed from a point z in it. As opposed to notions using Euclidean distance (say, the radius of the largest inscribed disk with center z), this notion is well-suited to use in complex analysis, in particular in conformal maps and conformal geometry.

A closely related notion is the transfinite diameter or (logarithmic) capacity of a compact simply connected set D, which can be considered as the inverse of the conformal radius of the complement E = Dc viewed from infinity.

Definition

Given a simply connected domain D ⊂ C, and a point z ∈ D, it then follows from the Riemann mapping theorem that there exists a unique conformal homeomorphism f : D → D onto the open unit disk (usually referred to as the uniformizing map) with f(z) = 0 ∈ D and f′(z) ∈ R+. The conformal radius of D from z is then defined as

rad⁡(z,D):=1f′(z).

The simplest example is that the conformal radius of the disk of radius r viewed from its center is also r, shown by the uniformizing map x ↦ x/r. See below for more examples.

One reason for the usefulness of this notion is that it behaves well under conformal maps: if φ: D → D′ is a conformal bijection and z in D, then rad⁡(φ(z),D′)=|φ′(z)|rad⁡(z,D).

The conformal radius can also be expressed as exp⁡(ξx(x)) where ξx(y) is the harmonic extension of log⁡(|x−y|) from ∂D to D.

A special case: the upper-half plane

Let K ⊂ H be a subset of the upper half-plane such that D:= H\K is connected and simply connected, and let z ∈ D be a point. (This is a usual scenario, say, in the Schramm–Loewner evolution). By the Riemann mapping theorem, there is a conformal bijection g: D → H. Then, for any such map g, a simple computation gives that

rad⁡(z,D)=2Im⁡(g(z))|g′(z)|.

For example, when K = ∅ and z = i, then g can be the identity map, and we get rad(i, H) = 2. Checking that this agrees with the original definition: the uniformizing map f : H → D is

f(z)=iz−iz+i,

and then the derivative can be easily calculated.

Relation to inradius

That it is a good measure of radius is shown by the following immediate consequence of the Schwarz lemma and the Koebe 1/4 theorem: for z ∈ D ⊂ C,

rad⁡(z,D)4≤dist⁡(z,∂D)≤rad⁡(z,D),

where dist(z, ∂D) denotes the Euclidean distance between z and the boundary of D, or in other words, the radius of the largest inscribed disk with center z.

Both inequalities are best possible:

The upper bound is clearly attained by taking D = D and z = 0.
The lower bound is attained by the following “slit domain”: D = C\R+ and z = −r ∈ R−. The square root map φ takes D onto the upper half-plane H, with φ(−r)=ir and derivative |φ′(−r)|=12r. The above formula for the upper half-plane gives rad⁡(ir,ℍ)=2r, and then the formula for transformation under conformal maps gives rad(−r, D) = 4r, while, of course, dist(−r, ∂D) = r.

Version from infinity: transfinite diameter and logarithmic capacity

Template:Main Template:Main When D ⊂ C is a connected, simply connected compact set, then its complement E = Dc is a connected, simply connected domain in the Riemann sphere that contains ∞,Template:Citation needed and one can define

rad⁡(∞,D):=1rad⁡(∞,E):=limz→∞f(z)z,

where f : C\D → E is the unique bijective conformal map with f(∞) = ∞ and that limit being positive real, i.e., the conformal map of the form

f(z)=c1z+c0+c−1z−1+⋯,c1∈𝐑+.

The coefficient c1 = rad(∞, D) equals the transfinite diameter and the (logarithmic) capacity of D; see Chapter 11 of Template:Harvtxt[1] and Template:Harvtxt.[2]

The coefficient c0 is called the conformal center of D. It can be shown to lie in the convex hull of D; moreover,

D⊆{z:|z−c0|≤2c1},

where the radius 2c1 is sharp for the straight line segment of length 4c1. See pages 12–13 and Chapter 11 of Template:Harvtxt.[1]

The Fekete, Chebyshev and modified Chebyshev constants

We define three other quantities that are equal to the transfinite diameter even though they are defined from a very different point of view. Let

d(z1,…,zk):=∏1≤i<j≤k|zi−zj|

denote the product of pairwise distances of the points z1,…,zk and let us define the following quantity for a compact set D ⊂ C:

dn(D):=supz1,…,zn∈Dd(z1,…,zn)1/(n2)

In other words, dn(D) is the supremum of the geometric mean of pairwise distances of n points in D. Since D is compact, this supremum is actually attained by a set of points. Any such n-point set is called a Fekete set.

The limit d(D):=limn→∞dn(D) exists and it is called the Fekete constant.

Now let 𝒫n denote the set of all monic polynomials of degree n in C[x], let 𝒬n denote the set of polynomials in 𝒫n with all zeros in D and let us define

μn(D):=infp∈𝒫nsupz∈D|p(z)| and μ~n(D):=infp∈𝒬nsupz∈D|p(z)|

Then the limits

μ(D):=limn→∞μn(D)1/n and μ(D):=limn→∞μ~n(D)1/n

exist and they are called the Chebyshev constant and modified Chebyshev constant, respectively. Michael Fekete and Gábor Szegő proved that these constants are equal.Template:Citation needed

Applications

The conformal radius is a very useful tool, e.g., when working with the Schramm–Loewner evolution. A beautiful instance can be found in Template:Harvtxt.[3]

References

Template:Sfn whitelist

Further reading