SLIDE 1
Finite Blaschke products and the construction of rational -inner - - PowerPoint PPT Presentation
Finite Blaschke products and the construction of rational -inner - - PowerPoint PPT Presentation
Finite Blaschke products and the construction of rational -inner functions Nicholas Young Leeds and Newcastle Universities Joint work with Jim Agler and Zinaida Lykova Bordeaux, 1st June 2015 Summary The symmetrised bidisc is the set def
SLIDE 2
SLIDE 3
Blaschke interpolation data
Consider points σ1, . . . , σ4, η1, . . . , η4 as in the diagram
η η η η
1 2 4 3
σ σ σ σ
4 3 1 2
and positive numbers ρ1, ρ2.
SLIDE 4
An interpolation problem for finite Blaschke products
Suppose given points σ1, . . . , σ4, η1, . . . , η4 and positive num- bers ρ1, ρ2 as on the preceding slide. Find if possible a Blaschke product ϕ of degree 4 with the properties ϕ(σj) = ηj for j = 1, . . . , 4 and Aϕ(σj) ≤ ρj for j = 1, 2, where Aϕ(eiθ) denotes the rate of change of the argument
- f ϕ(eiθ) with respect to θ.
SLIDE 5
The Pick matrix
associated with the foregoing interpolation problem is the matrix M =
- mij
4
i,j=1 where
mij =
ρj if i = j ≤ 2 1 − ¯ ηiηj 1 − ¯ σiσj
- therwise.
M ≥ 0 is necessary for the solvability of the Blaschke inter- polation problem.
SLIDE 6
Theorem
If the Pick matrix of the foregoing interpolation problem is positive definite then the problem has infinitely many so- lutions, which can be parametrised as follows. There are essentially unique polynomials a, b, c, d, of degree at most 4, such that the general solution of the problem is ϕζ(λ) = a(λ)ζ + b(λ) c(λ)ζ + d(λ) for λ ∈ D, where ζ ∈ T. There are explicit formulae for a, b, c, d, in which M−1 is prominent. See papers of Sarason, Georgijevi´ c, Bolotnikov-Dym, Chen- Hu.
SLIDE 7
Γ-inner functions
Recall that Γ def = {(z + w, zw) : |z| ≤ 1, |w| ≤ 1} ⊂ C2. The distinguished boundary of Γ is the ‘symmetrised torus’ bΓ = {(z + w, zw) : |z| = 1, |w| = 1}. A Γ-inner function is a map h ∈ Hol(D, Γ) whose boundary values lie in bΓ a.e. For each pair (ϕ, ψ) of (classical) inner functions, the func- tion (ϕ+ψ, ϕψ) is Γ-inner, but there are other Γ-inner func- tions, such as the degree 1 function (for β ∈ D) h(λ) = (β + ¯ βλ, λ). Describe all rational Γ-inner functions of degree n.
SLIDE 8
Royal nodes
The royal variety in Γ is the set R def = {(2λ, λ2) : |λ| ≤ 1} = {(s, p) ∈ Γ : s2 = 4p}. For any rational Γ-inner function h = (s, p), we say that λ ∈ D− is a royal node of h if h(λ) ∈ R, or equivalently, if s(λ)2 = 4p(λ). If h(λ) = (−2η, η2) then we say that η is the royal value corresponding to the royal node λ of h. Theorem If h = (s, p) is a rational Γ-inner function and p has degree n then h has exactly n royal nodes in D−, counted according to multiplicity.
SLIDE 9
Blaschke interpolation data again
The points σ1, . . . , σ4, η1, . . . , η4 are as in the diagram
η η η η
1 2 4 3
σ σ σ σ
4 3 1 2
and ρ1, ρ2 are positive numbers.
SLIDE 10
Γ-inner functions with prescribed royal nodes and values
Given Blaschke interpolation data σ1, . . . , σ4, η1, . . . , η4, ρ1, ρ2 as above, find a rational Γ-inner function h = (s, p) of degree 4 with royal nodes σ1, . . . , σ4 and corresponding royal values η1, . . . , η4 such that Ap(σj) = 2ρj, for j = 1, 2. Call these conditions (C).
SLIDE 11
An algorithm
- 1. Form the Pick matrix M corresponding to the Blaschke
interpolation data. If M is not positive definite then there is no Γ-inner function satisfying (C). Otherwise, let xλ
def
=
1 1−¯ σ1λ
. . .
1 1−¯ σ4λ
,
yλ
def
=
¯ η1 1−¯ σ1λ
. . .
¯ η4 1−¯ σ4λ
for λ ∈ D−.
- 2. Choose a point τ ∈ T \ {σ1, σ2}.
- 3. Find (s0, p0) ∈ bΓ such that
s0
- xλ, M−1xτ
- +
- yλ, M−1yτ
- +2
- xλ, M−1yτ
- +2p0
- yλ, M−1xτ
- is zero for all λ ∈ D.
SLIDE 12
Algorithm continued
- 4. If there is no (s0, p0) with these properties then there is
no Γ-inner function satisfying (C). Otherwise define polyno- mials a, b, c, d by explicit formulae in terms of σj, M, τ, xλ, yλ. (For example a(λ) =
4
- j=1
1 − ¯ σjλ 1 − ¯ σjτ
- 1 − (1 − ¯
τλ)
- xλ, M−1xτ
- .)
- 5. The function h = (s, p) is a Γ-inner function of degree 4
satisfying (C), where s = 22p0c − s0d s0c − 2d , p = −2p0a + s0b s0c − 2d .
SLIDE 13