Papers
Topics
Authors
Recent
2000 character limit reached

Dirichlet–Drury–Arveson Spaces

Updated 29 November 2025
  • Dirichlet–Drury–Arveson-type spaces are reproducing kernel Hilbert/Banach spaces of holomorphic functions defined via weighted sequence norms that generalize classical function spaces.
  • These spaces leverage complete Pick kernels and universal embeddings to connect finite and infinite-dimensional domains with practical applications in operator theory.
  • They bridge classical one-variable spaces with multi-variable and Dirichlet series contexts, offering insights into cyclicity, multiplier algebras, and spectral analysis.

A Dirichlet–Drury–Arveson-type space is a reproducing kernel Hilbert or Banach space of holomorphic functions on a multi-variable domain (typically a polydisc or an infinite-dimensional unit ball) defined via weighted sequence spaces, generalizing classical Dirichlet, Drury–Arveson, Hardy, and Besov–Sobolev spaces. These spaces are pivotal in function theory, operator theory, and the paper of Hilbert spaces of Dirichlet series, featuring intricate relationships with complete Pick kernels and multiplier algebras. The infinite-dimensional context, cyclicity properties, and universality phenomena differentiate the Dirichlet–Drury–Arveson-type settings from their classical finite-dimensional analogs.

1. Definition and Kernel Structure

Let N1N\geq1 (possibly infinite) and fix non-negative parameters s,t0s,t\geq0. For multi-indices α=(α1,,αN)Z+N\alpha=(\alpha_1,\dots,\alpha_N)\in\mathbb{Z}_+^N, set α=α1++αN|\alpha|=\alpha_1+\cdots+\alpha_N and α!=α1!αN!\alpha!=\alpha_1!\cdots\alpha_N!. Define the sequence weight

wα(s,t)=(α!α!)t(1+α)s.w_\alpha^{(s,t)} = \left(\frac{\alpha!}{|\alpha|!}\right)^t\cdot (1+|\alpha|)^s.

For 1p<1\leq p<\infty, the Banach space p(Z+N,w(s,t))\ell^p(\mathbb{Z}_+^N, w^{(s,t)}) consists of all x=(xα)x=(x_\alpha) with

xp=αxαp(wα(s,t))p<.\|x\|^p = \sum_\alpha |x_\alpha|^p (w_\alpha^{(s,t)})^p < \infty.

Under the Bohr (Fourier–Bohr) transform, xBx(ζ)=αxαζαx\mapsto Bx(\zeta) = \sum_{\alpha} x_\alpha \zeta^\alpha defines holomorphic functions on the Reinhardt domain DXND_{X'}^N, where XX' is the Köthe dual; for N=N=\infty, DXD_{X'}^\infty is the polydisc or infinite unit ball depending on (s,t,p)(s,t,p) (Nikolski, 22 Nov 2025).

The corresponding reproducing kernel (for p=2p=2) is

K(ζ,η)=αZ+Nζαηα[wα(s,t)]2.K(\zeta,\eta) = \sum_{\alpha\in\mathbb{Z}_+^N} \frac{\zeta^\alpha \overline{\eta}^\alpha}{[w_\alpha^{(s,t)}]^2}.

For (s=0,t=0)(s=0, t=0) and N=1N=1 this reduces to the Hardy kernel K(z,w)=1/(1zw)K(z,w)=1/(1-z\overline{w}). For the Drury–Arveson case (s=0,t=1/2)(s=0, t=1/2) it yields

K(ζ,η)=n=0ζ,ηn=11ζ,ηK(\zeta,\eta) = \sum_{n=0}^\infty \langle\zeta,\eta\rangle^n = \frac{1}{1-\langle\zeta,\eta\rangle}

on the unit ball B2NB_2^N (Nikolski, 22 Nov 2025).

2. Complete Pick Spaces and Universal Embedding

A central principle is the universality of the Drury–Arveson space Hd2H^2_d for complete Pick kernels: every RKHS with the complete Pick property can be realized as a quotient of some Hd2H^2_d via composition with a suitable map b:XBdb:X\to B_d. The Agler–McCarthy theorem guarantees that for a RKHS H\mathcal{H} with kernel KK having the Pick property, there exists dN{}d\in\mathbb{N}\cup\{\infty\} and b:XBdb:X\to B_d so that

K(z,w)=(1b(z),b(w))1K(z,w) = (1-\langle b(z), b(w)\rangle)^{-1}

(after absorbing weights into a renorming function) (Hartz, 2021).

For the classical Dirichlet space D\mathcal{D}, this embedding necessarily occurs with d=d=\infty: the explicit map

b(z)=(c1z,c2z2,c3z3,)2b(z) = (\sqrt{c_1}\, z,\, \sqrt{c_2}\, z^2,\, \sqrt{c_3}\, z^3,\dots) \in \ell^2

exhibits the infinite-dimensionality, since the Dirichlet kernel cannot be embedded into any finite BdB_d while preserving the surjectivity of the induced homomorphism on multiplier algebras (Hartz, 2021, Rochberg, 2016).

3. Dirichlet Series and Infinite-Dimensional Polydisks

Spaces of Dirichlet series with weighted norms (arising from anns\sum a_n n^{-s} expansions) can be identified with holomorphic function spaces on DD^\infty or BB_\infty via the Bohr lift: writing n=p1ν1pkνkn=p_1^{\nu_1} \cdots p_k^{\nu_k} and zν=z1ν1zkνkz^\nu=z_1^{\nu_1}\cdots z_k^{\nu_k}, the map f(s)=annsf~(z)=anzνf(s)=\sum a_n n^{-s} \mapsto \tilde{f}(z)=\sum a_n z^\nu is an isometric isomorphism (Olsen, 2010). The kernel structure and weight asymptotics determine local boundary behavior, Carleson measures, and interpolation properties, all governed by the order (α,β)(\alpha,\beta) of the partial sums W(X)=nXwnXα(logX)βW(X)=\sum_{n\leq X} w_n \sim X^\alpha (\log X)^\beta.

Infinite-variable analogs include:

  • Hardy on DD^\infty (wn1w_n\equiv 1)
  • Bergman on DD^\infty (coefficients wn=dβ+1(n)w_n=d_{\beta+1}(n))
  • Besov–Sobolev on BB_\infty (weights wn=w_n = multinomial coefficients)
  • Drury–Arveson on BB_\infty (kernel 1/(1z,ζ)1/(1-\langle z,\zeta\rangle)) (Olsen, 2010, Nikolski, 22 Nov 2025)

4. Multiplier Algebras and Quotient Structures

The multiplier algebra Mult(H)\mathrm{Mult}(\mathcal{H}) of a Dirichlet–Drury–Arveson-type space is tightly linked to its embedding dimension. For D\mathcal{D}, the induced homomorphism Φ:Mult(Hd2)Mult(D), φφb\Phi: \mathrm{Mult}(H^2_d)\to \mathrm{Mult}(\mathcal{D}),\ \varphi\mapsto \varphi\circ b, is a surjective, contractive (and in fact isometric) quotient map only for d=d=\infty. The finite-dimensional Drury–Arveson multiplier algebra Mult(Hd2)H(Bd)\mathrm{Mult}(H^2_d)\simeq H^\infty(B_d) is tractable and has maximal-ideal fibers over Bd\overline{B}_d; by contrast, Mult(D)\mathrm{Mult}(\mathcal{D}) inherits complexity (corona phenomena) from the non-metrizable βNN\beta \mathbb{N}\setminus \mathbb{N} component forced by d=d=\infty (Hartz, 2021).

More generally, for Dirichlet series RKHSs, if the embedding is f(s)=(b1n1s,b2n2s,)f(s)=(b_1 n_1^{-s}, b_2 n_2^{-s}, \ldots), then rational independence of the lognk\log n_k implies the closure of the image is all of BdB_d and the multiplier algebra is unitarily, isometrically isomorphic to Mult(Hd2)\mathrm{Mult}(H^2_d); otherwise, it is a proper quotient (McCarthy et al., 2015). Every complete Pick algebra is a quotient of such a Dirichlet–Drury–Arveson-type multiplier algebra.

5. Cyclicity, Dilated Systems, and Factorization

The cyclicity of dilated systems (i.e., when the family {Dnx:n1}\{D_n x:n\geq1\} spans the space) in Dirichlet–Drury–Arveson-type sequence lattices is closely linked to multiplicative structure, invertibility in associated power-series algebras, and the notion of dominating free term. Under the Bohr transform, dilations correspond to monomial multiplication in function spaces on DXD_{X'}^\infty or BpNB_{p'}^N. For totally multiplicative sequences, cyclicity is guaranteed by the extended Haar lemma, and functions with sufficiently large free term (in 1\ell^1 Banach algebras) are invertible and thus cyclic (Nikolski, 22 Nov 2025).

Polynomials with linear factorization—splitting into products of affine factors—are cyclic if they have no zeros on the spectrum. Cyclicity results vary with underlying domain and parameter values: in Drury–Arveson or Hilbertian Dirichlet-type cases, generic cyclicity may require deeper boundary conditions, especially in finite dimensions (Nikolski, 22 Nov 2025).

6. Operator Theory and Function-Theoretic Consequences

The infinite embedding dimension yields significant implications for operator models. The Dirichlet shift cannot be modeled by any finite commuting dd-tuple of weighted shifts on the ball, precluding finite-dimensional function-theoretic realizations for certain Dirichlet-type operators (Hartz, 2021, Rochberg, 2016). In Drury–Arveson spaces, the canonical Gleason problem admits contractive multi-operator solutions, and de Branges–Rovnyak-type subspaces may or may not be invariant under coordinate multipliers, depending on quasi-extremality—through the noncommutative Herglotz representation and GNS constructions (Jury, 2013).

Boundary behavior, spectral analysis, and Clark theory extend into the multivariable and infinite-variable context, requiring new operator-algebraic machinery, such as Cuntz–Toeplitz algebras and noncommutative Fantappiè transforms.

7. Connections to Classical Scales and Invariant Theory

Dirichlet–Drury–Arveson-type spaces interpolate between the classical Hardy, Bergman, Dirichlet, Besov–Sobolev, and de Branges–Rovnyak settings. The table below summarizes parameter regimes and space identification (adapted from (Olsen, 2010, Nikolski, 22 Nov 2025)):

(α,β)(\alpha, \beta) Weight W(X)W(X) Local Model Infinite-var. space
(0,0)(0,0) XX H2(C1/2)H^2(C_{1/2}) Dirichlet–Hardy H2(D)H^2(D^\infty)
(1,0)(1,0) X/lnXX/\ln X D0(C1/2)D_0(C_{1/2}) Drury–Arveson H2(B)H^2(B_\infty)
(1,β>0)(1,\beta>0) X(lnX)βX(\ln X)^\beta Dβ(C1/2)D_{-\beta}(C_{1/2}) Bergman Aβ(D)A_\beta(D^\infty)
(α>0,0)(\alpha>0,0) XαX^\alpha D1α(C1/2)D_{1-\alpha}(C_{1/2}) Besov–Sobolev B2α(B)B_2^\alpha(B_\infty)

This framework enables a unified treatment of cyclicity and multiplier algebra structure across domains and parameter regimes, showing the precise specialization to one-variable or multi-variable operator theory, and highlighting the essential difference between finite and infinite-dimensional spaces with respect to universality, metric geometry, and functional models (Olsen, 2010, McCarthy et al., 2015, Nikolski, 22 Nov 2025).

Slide Deck Streamline Icon: https://streamlinehq.com

Whiteboard

Forward Email Streamline Icon: https://streamlinehq.com

Follow Topic

Get notified by email when new papers are published related to Dirichlet--Drury--Arveson-type Spaces.