betaFeedback

WorksMathematics

Functions of Sp\boldsymbol{S}_p-almost commuting unitary operators and contractions. An extension of the Carey–Pincus trace formula

Abstract

In this paper we study functions of pairs of unitary operators and pairs of contractions with commutators in the Schatten-von Neumann class Sp\boldsymbol{S}_p. We obtain commutator estimates and prove that our functional calculus is multiplicative modulo Sp\boldsymbol{S}_p. On the other hand, we show that our commutator estimates do not extend to the case p>2p>2. We also obtain an extension of the Carey–Pincus trace formula. The main tool of this paper is a technique of triple operator integrals.

Introduction

In [3] we studied functions of almost commuting self-adjoint operators and obtained essential improvements of earlier results of [20]. Recall that operators TT and RR on Hilbert space are called almost commuting if their commutator [T,R][T,R] ≔\coloneqq TR−RTTR-RT belongs to the trace class S1S_1. In this paper we are going to study functions of almost commuting unitary operators and almost commuting contractions. In fact, we consider the more general case of pp-almost commuting unitary operators and pp-almost commuting contractions (operators TT and RR on Hilbert space are called pp-almost commuting if [T,R][T,R] belongs to the Schatten–von Neumann class SpS_p).

Note also that the case of pp-almost commuting self-adjoint operators was considered in [23].

Recall that the function φ(U,V)\varphi(U,V) of not necessarily commutingunitary operators UU and VV is defined as the double operator integral

φ(U,V)=∬T2φ(x,y) dEU(x) dEV(y),\varphi(U,V)=\iint_{\mathbb{T}^{2}}\varphi(x,y)\,dE_U(x)\,dE_V(y),

where EUE_U and EVE_V are the spectral measures of UU and VV.

Recall that the theory of double operator integrals of the form

∬Φ(x,y) dE1(x)Q dE2(y)(1)\iint\Phi(x,y)\,dE_1(x)Q\,dE_2(y) \tag*{(1)}

was developed by M.Sh. Birman and M.Z. Solomyak in [7]–[9]. Here E1E_1 and E2E_2 are spectral measures on Hilbert space, Φ\Phi is a bounded measurable function and QQ is a bounded linear operator on Hilbert space. We also refer the reader to [2] and [22] for more information on double operator integrals.

The maximal class of functions Φ\Phi for which the double operator integral (1.1) can be defined in a natural way as a bounded linear operator for an arbitrary bounded operator QQ is called the class of Schur multipliers with respect to E1E_1 and E2E_2. This class is denoted by ME1,E2\mathfrak{M}_{E_1,E_2}.

It is well known (see [18], [2] and [6]) that Φ∈ME1,E2\Phi\in\mathfrak{M}_{E_1,E_2} if and only if Φ\Phi belongs to the Haagerup tensor product LE1∞⊗hLE2∞L^\infty_{E_1} \otimes_{\mathrm h} L^\infty_{E_2} of the spaces LE1∞L^\infty_{E_1} and LE2∞L^\infty_{E_2}, i.e. Φ\Phi admits a representation

Φ(x,y)=∑nun(x)vn(y),\Phi(x,y)=\sum_n u_n(x)v_n(y),

with functions unu_n in LE1∞L^\infty_{E_1} and vnv_n in LE2∞L^\infty_{E_2} satisfying the conditions

∑n∣un∣2∈LE1∞and∑n∣vn∣2∈LE2∞.\sum_n |u_n|^2 \in L^\infty_{E_1} \quad\text{and}\quad \sum_n |v_n|^2 \in L^\infty_{E_2}.

Note that in the same way one can define the Haagerup tensor product of subspaces of L∞L^\infty.

It was observed in [1] that if φ\varphi is a function on T2\mathbb{T}^2 that belongs to the homogeneous Besov class B∞,11(T2)B^1_{\infty,1}(\mathbb{T}^2) (see § 2), then φ∈LE1∞⊗hLE2∞\varphi\in L^\infty_{E_1} \otimes_{\mathrm h} L^\infty_{E_2} for arbitrary unitary operators UU and VV which allows one to define the functional calculus

φ↦φ(U,V),φ∈B∞,11(T2).\varphi\mapsto\varphi(U,V),\qquad\varphi\in B^1_{\infty,1}(\mathbb{T}^2).

It was shown in [1] (see Theorem 12.4) that this functional calculus satisfies a Lipschitz type estimate in the Schatten–von Neumann norm of SpS_p for p∈[1,2]p \in[1,2], i.e. the following inequality holds

∥φ(U1,V1)−φ(U2,V2)∥Sp≤const⁡ ∥φ∥B∞,11max⁡{∥U2−U1∥Sp,∥V2−V1∥Sp}.(2)\lVert\varphi(U_1,V_1)-\varphi(U_2,V_2)\rVert_{S_p} \leq\operatorname{const}\,\lVert\varphi\rVert_{B^1_{\infty,1}} \max\{\lVert U_2-U_1\rVert_{S_p},\lVert V_2-V_1\rVert_{S_p}\}. \tag*{(2)}

whenever U1U_1, V1V_1, U2U_2 and V2V_2 are bounded self-adjoint operators such that U2−U1∈SpU_2-U_1 \in S_p and V2−V1∈SpV_2-V_1 \in S_p. We refer the reader to [12] for the definition and basic properties of Schatten–von Neumann classes.

On the other hand, it was proved in [5] (see Theorem 7.3) that inequality (1.2) is false in the norm of SpS_p for p>2p>2 and it is also false in the operator norm.

In this paper we consider the functional calculus φ↦φ(U,V)\varphi\mapsto\varphi(U,V), φ∈B∞,11(T2)\varphi\in B^1_{\infty,1}(\mathbb{T}^2), for pp-almost commuting unitary operators UU and VV, i.e. unitary operators with commutators in the Schatten–von Neumann class SpS_p.

We show that in § 4 that in the case p∈[1,2]p \in[1,2] the calculus φ↦φ(U,V)\varphi\mapsto\varphi(U,V), φ∈B∞,11(T2)\varphi\in B^1_{\infty,1}(\mathbb{T}^2), admits a commutator Lipschitz estimate in the Schatten–von Neumann norm SpS_p. However, we establish in § 7 that the situation completely changes in the norm of SpS_p for p>2p>2 as well as in the operator norm. Also, we show that for p∈[1,2]p \in[1,2], the calculus φ↦φ(U,V)\varphi\mapsto\varphi(U,V) is multiplicative modulo SpS_p.

In § 5 we establish an extension of the Carey–Pincus trace formula [10] which is an analogue of the Helton–Howe trace formula [13].

In § 6 we obtain analogs of the results of § 4 for functions of contractions while in § 7 we show that the commutator estimate obtained in § 4 do not extend to the case p>2p>2.

In § 2 we define the Besov space B∞,11(T2)B^1_{\infty,1}(\mathbb{T}^2) which plays a very important role in this paper. Finally, we give a detailed introduction in triple operator integrals in § 3.

The Besov space B∞,11(T2)B^1_{\infty,1}(\mathbb{T}^2) of periodic function

In this section we define the Besov class B∞,11(T2)B^1_{\infty,1}(\mathbb{T}^2) of two variables which will play an important role in this paper.

Let ww be an infinitely differentiable function on R\mathbb{R} such that

w≥0,supp⁡w⊂[12,2],andw(s)=1−w(s2)for s∈[1,2](3)w \ge0,\qquad\operatorname{supp} w \subset\left[\frac{1}{2},2\right],\qquad\text{and}\quad w(s)=1-w\left(\frac{s}{2}\right)\quad\text{for }s\in[1,2] \tag*{(3)}

and define the trigonometric polynomials WnW_n, n≥0n\ge0, by

Wn(ζ)≔∑j∈Z2w(∣j∣2n)ζj,n≥1,W0(ζ)≔∑{j:∣j∣≤1}ζj,(4)W_n(\zeta)\coloneqq\sum_{j\in\mathbb{Z}^2}w\left(\frac{|j|}{2^n}\right)\zeta^j,\qquad n\ge1,\qquad W_0(\zeta)\coloneqq\sum_{\{j:|j|\le1\}}\zeta^j, \tag*{(4)}

where

ζ=(ζ1,ζ2)∈Td,j=(j1,j2),and∣j∣=(∣j1∣2+∣j2∣2)1/2.\zeta=(\zeta_1,\zeta_2)\in\mathbb{T}^d,\qquad j=(j_1,j_2),\qquad\text{and}\quad|j|=(|j_1|^2+|j_2|^2)^{1/2}.

For a distribution ff on T2\mathbb{T}^2 we put

fn=f∗Wn,n≥0,f_n=f*W_n,\qquad n\ge0,

and we say that ff belongs to the Besov class Bp,qs(T2)B^s_{p,q}(\mathbb{T}^2), s>0s>0, 1≤p,q≤∞1\le p,q\le\infty, if

{2ns∥fn∥Lp}n≥0∈ℓq.(5)\left\{2^{ns}\|f_n\|_{L^p}\right\}_{n\ge0}\in\ell^q. \tag*{(5)}

We refer the reader to [2], [17] and [23] for the definition and basic properties of Besov spaces.

Triple operator integrals

Both in [1] and [3] an important role was played by triple operator integrals, i.e. expressions of the form

∭Φ(x,y,z) dE1(x)Q1 dE2(y)Q2 dE3(z).(6)\iiint\Phi(x,y,z)\,dE_1(x)Q_1\,dE_2(y)Q_2\,dE_3(z). \tag*{(6)}

Here E1E_1, E2E_2 and E3E_3 are spectral measures on Hilbert space, Q1Q_1 and Q2Q_2 are bounded linear operators on Hilbert space and Φ\Phi is a bounded measurable function. We call Φ\Phi the symbol of the triple operator integral (6).

As in the case of double operator integrals, for such triple operator integrals to make sense, the function Φ\Phi and the operators Q1Q_1 and Q2Q_2 have to satisfy certain conditions. We mention here several approaches to triple operator integrals.

In [20] triple operator integrals of the form (6) were defined for functions Φ\Phi in the integral projective tensor product LE1∞⊗^iLE2∞⊗^iLE3∞L^\infty_{E_1}\widehat{\otimes}_{i}L^\infty_{E_2}\widehat{\otimes}_{i}L^\infty_{E_3}, see [20] for the definition. It turns out that if Φ∈LE1∞⊗^iLE2∞⊗^iLE3∞\Phi\in L^\infty_{E_1}\widehat{\otimes}_{i}L^\infty_{E_2}\widehat{\otimes}_{i}L^\infty_{E_3}, Q1Q_1 and Q2Q_2 are bounded operators, then the triple operator integral in (6) determines a bounded linear operator. Also, for Φ∈LE1∞⊗^iLE2∞⊗^iLE3∞\Phi\in L^\infty_{E_1}\widehat{\otimes}_{i}L^\infty_{E_2}\widehat{\otimes}_{i}L^\infty_{E_3}, the conditions that Q1∈SpQ_1\in S_p and Q2∈SqQ_2\in S_q with 1/p+1/q≤11/p+1/q\le1 ensure that the triple operator integral in (6) belongs to the Schatten–von Neumann class SrS_r, where 1/r=1/p+1/q1/r=1/p+1/q.

Let us also mention an earlier definition of triple, or even multiple, operator integrals given in [St] as well as the definition of multiple operator integrals for Hilbert–Schmidt operators given in [16].

We dwell on the next two approaches in more detail. In [14] the authors considered triple operator integrals of the form (3.1) in the case when Φ\Phi belongs to the Haagerup tensor product LE1∞⊗hLE2∞⊗hLE3∞L^\infty_{E_1} \otimes_{\mathrm{h}} L^\infty_{E_2} \otimes_{\mathrm{h}} L^\infty_{E_3} of the L∞L^\infty spaces which consists of functions Φ\Phi of the form

Φ(x,y,z)=∑j,k≥0αj(x1)βjk(x2)γk(x3)\Phi(x,y,z)=\sum_{j,k\geq0}\alpha_j(x_1)\beta_{jk}(x_2)\gamma_k(x_3)

where {αj}j≥0\{\alpha_j\}_{j\geq0}, {γk}k≥0∈L∞(ℓ2)\{\gamma_k\}_{k\geq0}\in L^\infty(\ell^2), {βjk}j,k≥0∈L∞(B)\{\beta_{jk}\}_{j,k\geq0}\in L^\infty(\mathcal{B}), where B\mathcal{B} is the space of matrices that induce bounded linear operators on $\ell^2 equipped with the operator norm. For such functions Φ\Phi the triple operator integral in (3.1) can be expressed in the following way:

∭Φ(x,y,z) dE1(x)Q1 dE2(y)Q2 dE3(z)=∑j,k≥0(∫αj dE1)Q1(∫βjk dE2)Q2(∫γk dE3)=lim⁡M,N→∞∑j=0N∑k=0M(∫αj dE1)Q1(∫βjk dE2)Q2(∫γk dE3),\begin{aligned} \iiint\Phi(x,y,z)\,\mathrm{d}E_1(x)Q_1\,\mathrm{d}E_2(y)Q_2\,\mathrm{d}E_3(z) \\ &=\sum_{j,k\geq0}\left(\int\alpha_j\,\mathrm{d}E_1\right)Q_1\left(\int\beta_{jk}\,\mathrm{d}E_2\right)Q_2\left(\int\gamma_k\,\mathrm{d}E_3\right) \\ &=\lim_{M,N\to\infty}\sum_{j=0}^{N}\sum_{k=0}^{M}\left(\int\alpha_j\,\mathrm{d}E_1\right)Q_1\left(\int\beta_{jk}\,\mathrm{d}E_2\right)Q_2\left(\int\gamma_k\,\mathrm{d}E_3\right), \end{aligned}

where the limit on the right exists in the weak operator topology. Also, the triple operator integral in (3.1) can be expressed as the product of the block operator matrices

∭Φ(x,y,z) dE1(x)Q1 dE2(y)Q2 dE3(z)=(A1A2A3⋯ )(B11B12B13⋯B21B22B23⋯B31B32B33⋯⋮⋮⋮⋱)(Γ1Γ2Γ3⋮),\begin{aligned} \iiint\Phi(x,y,z)\,\mathrm{d}E_1(x)Q_1\,\mathrm{d}E_2(y)Q_2\,\mathrm{d}E_3(z) \\ &=(A_1\quad A_2\quad A_3\quad\cdots) \begin{pmatrix} B_{11} & B_{12} & B_{13} & \cdots\\ B_{21} & B_{22} & B_{23} & \cdots\\ B_{31} & B_{32} & B_{33} & \cdots\\ \vdots& \vdots& \vdots& \ddots \end{pmatrix} \begin{pmatrix} \Gamma_1\\ \Gamma_2\\ \Gamma_3\\ \vdots \end{pmatrix}, \end{aligned}

where

Aj≔(∫αj dE1)Q1,Bjk≔∫βjk dE2andΓk≔Q2∫γk dE3.A_j \coloneqq\left(\int\alpha_j\,\mathrm{d}E_1\right)Q_1,\quad B_{jk}\coloneqq\int\beta_{jk}\,\mathrm{d}E_2\quad\text{and}\quad\Gamma_k\coloneqq Q_2\int\gamma_k\,\mathrm{d}E_3.

All three block operator matrices on the right-hand side determine bounded linear operators, see [4].

It was established in [4] that if Φ∈LE1∞⊗hLE2∞⊗hLE3∞\Phi\in L^\infty_{E_1}\otimes_{\mathrm{h}}L^\infty_{E_2}\otimes_{\mathrm{h}}L^\infty_{E_3}, Q1∈SpQ_1\in S_p, Q2∈SqQ_2\in S_q with p,q∈[2,∞]p,q\in[2,\infty], then the triple operator integral in (3.1) belongs to SrS_r, where 1/r=1/p+1/q1/r=1/p+1/q. Here and in what follows for convenience, by T∈S∞T\in S_\infty we mean that TT is a bounded linear operator.

We proceed now to the most important approach for us to triple operator integrals. This approach was offered in [1] and was slightly refined in [4]. It allows us to define triple operator integrals of the form (3.1) for functions Φ\Phi in Haagerup-like tensor products of L∞L^\infty spaces.

We say that a measurable function Φ\Phi of three variables belongs to the Haagerup-like tensor product L∞(E1)⊗hL∞(E2)⊗hL∞(E3)L^\infty(E_1)\mathbin{\otimes_{\mathrm{h}}}L^\infty(E_2)\mathbin{\otimes_{\mathrm{h}}}L^\infty(E_3) of the first kind if it admits a representation

Φ(x,y,z)=∑j,k≥0αj(x)βk(y)γjk(z)(7)\Phi(x,y,z)=\sum_{j,k\geq0}\alpha_j(x)\beta_k(y)\gamma_{jk}(z) \tag*{(7)}

with {αj}j≥0\{\alpha_j\}_{j\geq0}, {βk}k≥0∈L∞(ℓ2)\{\beta_k\}_{k\geq0}\in L^\infty(\ell^2) and {γjk}j,k≥0∈L∞(B)\{\gamma_{jk}\}_{j,k\geq0}\in L^\infty(\mathcal{B}). We equip the space L∞(E1)⊗hL∞(E2)⊗hL∞(E3)L^\infty(E_1)\mathbin{\otimes_{\mathrm{h}}}L^\infty(E_2)\mathbin{\otimes_{\mathrm{h}}}L^\infty(E_3) with the norm

∥Φ∥L∞⊗hL∞⊗hL∞≔inf⁡∥{αj}j≥0∥L∞(ℓ2)∥{βk}k≥0∥L∞(ℓ2)∥{γjk}j,k≥0∥L∞(B),\lVert\Phi\rVert_{L^\infty\otimes_{\mathrm{h}}L^\infty\otimes_{\mathrm{h}}L^\infty}\coloneqq\inf\lVert\{\alpha_j\}_{j\geq0}\rVert_{L^\infty(\ell^2)}\lVert\{\beta_k\}_{k\geq0}\rVert_{L^\infty(\ell^2)}\lVert\{\gamma_{jk}\}_{j,k\geq0}\rVert_{L^\infty(\mathcal{B})},

the infimum being taken over all representations of the form (3.2).

Let us now define triple operator integrals of the first gender with symbol Φ\Phi in the tensor product L∞(E1)⊗hL∞(E2)⊗hL∞(E3)L^\infty(E_1)\mathbin{\otimes_{\mathrm{h}}}L^\infty(E_2)\mathbin{\otimes_{\mathrm{h}}}L^\infty(E_3).

Suppose that 1≤p≤21\leq p\leq2. For Φ∈L∞(E1)⊗hL∞(E2)⊗hL∞(E3)\Phi\in L^\infty(E_1)\mathbin{\otimes_{\mathrm{h}}}L^\infty(E_2)\mathbin{\otimes_{\mathrm{h}}}L^\infty(E_3), for a bounded linear operator TT, and for an operator RR of class Sp\mathcal{S}_p, we define the triple operator integral

W=∭Φ(x,y,z) dE1(x)T dE2(y)R dE3(z)W=\iiint\Phi(x,y,z)\,\mathrm{d}E_1(x)T\,\mathrm{d}E_2(y)R\,\mathrm{d}E_3(z)

of the first gender as the following continuous linear functional on Sp′\mathcal{S}_{p'}, 1/p+1/p′=11/p+1/p'=1 (on the class of compact operators in the case p=1p=1):

Q↦trace⁡((∭Φ(x,y,z) dE2(y)R dE3(z)Q dE1(x))T).Q\mapsto\operatorname{trace}\left(\left(\iiint\Phi(x,y,z)\,\mathrm{d}E_2(y)R\,\mathrm{d}E_3(z)Q\,\mathrm{d}E_1(x)\right)T\right).

It follows easily from the above Schatten–von Neumann properties of triple operator integrals with symbols in the Haagerup tensor product L∞⊗hL∞⊗hL∞L^\infty\mathbin{\otimes_{\mathrm{h}}}L^\infty\mathbin{\otimes_{\mathrm{h}}}L^\infty that W∈SpW\in\mathcal{S}_p and

∥W∥Sp≤∥Ψ∥L∞⊗hL∞⊗hL∞∥T∥Sp∥R∥,1≤p≤2,(8)\lVert W\rVert_{\mathcal{S}_p}\leq\lVert\Psi\rVert_{L^\infty\otimes_{\mathrm{h}}L^\infty\otimes_{\mathrm{h}}L^\infty}\lVert T\rVert_{\mathcal{S}_p}\lVert R\rVert,\qquad1\leq p\leq2, \tag*{(8)}

see [1] and [4]. Note that a considerably more general result on Schatten–von Neumann properties of triple operator integrals of the first gender is given in Theorem 5.1 of [4].

Let us proceed now to Haagerup tensor products of the second kind and triple operator integrals of the second gender.

We say that a function Φ\Phi belongs to the Haagerup-like tensor product L∞(E1)⊗hL∞(E2)⊗hL∞(E3)L^\infty(E_1)\mathbin{\otimes_{\mathrm{h}}}L^\infty(E_2)\mathbin{\otimes_{\mathrm{h}}}L^\infty(E_3) of the second kind if Φ\Phi admits a representation

Φ(x,y,z)=∑j,k≥0αjk(x)βj(y)γk(z)(9)\Phi(x,y,z)=\sum_{j,k\geq0}\alpha_{jk}(x)\beta_j(y)\gamma_k(z) \tag*{(9)}

where {βj}j≥0\{\beta_j\}_{j\geq0}, {γk}k≥0∈L∞(ℓ2)\{\gamma_k\}_{k\geq0}\in L^\infty(\ell^2), {αjk}j,k≥0∈L∞(B)\{\alpha_{jk}\}_{j,k\geq0}\in L^\infty(\mathcal{B}). The norm of Φ\Phi in the space L∞⊗hL∞⊗hL∞L^\infty\mathbin{\otimes_{\mathrm{h}}}L^\infty\mathbin{\otimes_{\mathrm{h}}}L^\infty is defined by

∥Φ∥L∞⊗hL∞⊗hL∞≔inf⁡∥{αjk}j,k≥0∥L∞(B)∥{βj}j≥0∥L∞(ℓ2)∥{γk}k≥0∥L∞(ℓ2),\lVert\Phi\rVert_{L^\infty\otimes_{\mathrm{h}}L^\infty\otimes_{\mathrm{h}}L^\infty}\coloneqq\inf\lVert\{\alpha_{jk}\}_{j,k\geq0}\rVert_{L^\infty(\mathcal{B})}\lVert\{\beta_j\}_{j\geq0}\rVert_{L^\infty(\ell^2)}\lVert\{\gamma_k\}_{k\geq0}\rVert_{L^\infty(\ell^2)},

the infimum being taken over all representations of the form (3.4).

We proceed now to the definition of triple operator integrals of the second gender with integrand in the tensor product L∞(E1)⊗hL∞(E2)⊗hL∞(E3)L^\infty(E_1)\otimes_h L^\infty(E_2)\otimes_h L^\infty(E_3).

Suppose that Φ∈L∞(E1)⊗hL∞(E2)⊗hL∞(E3)\Phi\in L^\infty(E_1)\otimes_h L^\infty(E_2)\otimes_h L^\infty(E_3), TT is a bounded linear operator, and R∈SpR\in S_p, 1≤p≤21\le p\le2. The continuous linear functional

Q↦trace⁡((∭Φ(x,y,z) dE3(z)Q dE1(x)T dE2(y))R)Q\mapsto\operatorname{trace}\left(\left(\iiint\Phi(x,y,z)\,dE_3(z)Q\,dE_1(x)T\,dE_2(y)\right)R\right)

on the class Sp′S_{p'} (on the class of compact operators in the case p=1p=1) determines an operator WW of class SpS_p, which we call the triple operator integral

W=∭Φ(x,y,z) dE1(x)T dE2(y)R dE3(z)W=\iiint\Phi(x,y,z)\,dE_1(x)T\,dE_2(y)R\,dE_3(z)

of the second gender.

It is easy to see that

∥W∥Sp≤∥Φ∥L∞⊗hL∞⊗hL∞∥T∥⋅∥R∥Sp,(10)\lVert W\rVert_{S_p}\le\lVert\Phi\rVert_{L^\infty\otimes_hL^\infty\otimes_hL^\infty}\lVert T\rVert\cdot\lVert R\rVert_{S_p}, \tag*{(10)}

see [1] and [4].

A considerably more general result on Schatten–von Neumann properties of triple operator integrals of the second gender is given in Theorem 5.2 of [4].

Functions of pp-almost commuting unitary operators

In this section we obtain commutator estimates for the calculus φ↦φ(U,V)\varphi\mapsto\varphi(U,V), φ∈B∞,1(T2)\varphi\in B_{\infty,1}(\mathbb{T}^2) in Schatten–von Neumann norms SpS_p, 1≤p≤21\le p\le2, and show that this calculus is multiplicative modulo SpS_p.

With a function φ\varphi in B∞,1(T2)B_{\infty,1}(\mathbb{T}^2) we associate two divided differences defined on T2\mathbb{T}^2 by

(D[1]φ)(ζ1,ζ2,τ)≔φ(ζ1,τ)−φ(ζ2,τ)ζ1−ζ2and(D[2]φ)(ζ,τ1,τ2)≔φ(ζ,τ1)−φ(ζ,τ2)τ1−τ2.(\mathfrak{D}^{[1]}\varphi)(\zeta_1,\zeta_2,\tau)\coloneqq\frac{\varphi(\zeta_1,\tau)-\varphi(\zeta_2,\tau)}{\zeta_1-\zeta_2} \qquad\text{and}\qquad (\mathfrak{D}^{[2]}\varphi)(\zeta,\tau_1,\tau_2)\coloneqq\frac{\varphi(\zeta,\tau_1)-\varphi(\zeta,\tau_2)}{\tau_1-\tau_2}.

It was proved in [1], Theorem 12.3 that for φ∈B∞,1(T2)\varphi\in B_{\infty,1}(\mathbb{T}^2), D[1]φ∈L∞⊗hL∞⊗hL∞\mathfrak{D}^{[1]}\varphi\in L^\infty\otimes_hL^\infty\otimes_hL^\infty, D[2]φ∈L∞⊗hL∞⊗hL∞\mathfrak{D}^{[2]}\varphi\in L^\infty\otimes_hL^\infty\otimes_hL^\infty,

∥D[1]φ∥L∞⊗hL∞⊗hL∞≤const⁡∥φ∥B∞,11and∥D[2]φ∥L∞⊗hL∞⊗hL∞≤const⁡∥φ∥B∞,11,\lVert\mathfrak{D}^{[1]}\varphi\rVert_{L^\infty\otimes_hL^\infty\otimes_hL^\infty}\le\operatorname{const}\lVert\varphi\rVert_{B^1_{\infty,1}} \qquad\text{and}\qquad \lVert\mathfrak{D}^{[2]}\varphi\rVert_{L^\infty\otimes_hL^\infty\otimes_hL^\infty}\le\operatorname{const}\lVert\varphi\rVert_{B^1_{\infty,1}},

where by L∞L^\infty we can mean the L∞L^\infty space with respect to arbitrary Borel measures.

Theorem 4.1. Let 1≤p≤21\le p\le2 and φ∈B∞,1(T2)\varphi\in B_{\infty,1}(\mathbb{T}^2). Let UU and VV be unitary operators and let QQ be a bounded linear operator such that [U,Q]∈Sp[U,Q]\in S_p and [V,Q]∈Sp[V,Q]\in S_p. Suppose that φ\varphi is a function in the Besov class B∞,11(T2)B^1_{\infty,1}(\mathbb{T}^2). Then [φ(U,V),Q]∈Sp[\varphi(U,V),Q]\in S_p,

[φ(U,V),Q]=∭φ(ζ,τ1)−φ(ζ,τ2)τ1−τ2 dEU(ζ) dEV(τ1)[V,Q] dEV(τ2)+∭φ(ζ1,τ)−φ(ζ2,τ)ζ1−ζ2 dEU(ζ1)[U,Q] dEU(ζ2) dEV(τ).(11)[\varphi(U,V),Q]=\iiint\frac{\varphi(\zeta,\tau_1)-\varphi(\zeta,\tau_2)}{\tau_1-\tau_2}\,dE_U(\zeta)\,dE_V(\tau_1)[V,Q]\,dE_V(\tau_2) +\iiint\frac{\varphi(\zeta_1,\tau)-\varphi(\zeta_2,\tau)}{\zeta_1-\zeta_2}\,dE_U(\zeta_1)[U,Q]\,dE_U(\zeta_2)\,dE_V(\tau). \tag*{(11)}

:::`

and

∥[φ(U,V),Q]∥Sp≤const⁡∥φ∥B∞,11(T2)(∥[U,Q]∥Sp+∥[V,Q]∥Sp).(12)\lVert[\varphi(U,V),Q]\rVert_{\mathcal{S}_p}\leq\operatorname{const}\lVert\varphi\rVert_{B^1_{\infty,1}(\mathbb{T}^2)}\left(\lVert[U,Q]\rVert_{\mathcal{S}_p}+\lVert[V,Q]\rVert_{\mathcal{S}_p}\right). \tag*{(12)}

Proof. The equality (11) can be proved by straightforward calculations for φ(ζ,τ)=ζmτn\varphi(\zeta,\tau)=\zeta^m\tau^n, where m,n∈Zm,n\in\mathbb{Z}. Hence, (11) and (12) hold for the trigonometric polynomials φ\varphi on T2\mathbb{T}^2.

Suppose now that φ\varphi is an arbitrary function in B∞,11(T2)B^1_{\infty,1}(\mathbb{T}^2). Then D[1]φ∈L∞⊗hL∞⊗hL∞D^{[1]}\varphi\in L^\infty\otimes_{\mathrm{h}}L^\infty\otimes_{\mathrm{h}}L^\infty and D[2]φ∈L∞⊗hL∞⊗hL∞D^{[2]}\varphi\in L^\infty\otimes_{\mathrm{h}}L^\infty\otimes_{\mathrm{h}}L^\infty by Theorem 12.3 in [1]. Hence, the integrals in (11) are well defined. The function φ∈B∞,11(T2)\varphi\in B^1_{\infty,1}(\mathbb{T}^2) can be can be represented in the form

φ=∑n≥0φn,\varphi=\sum_{n\geq0}\varphi_n,

where φn:=φ∗Wn\varphi_n\mathrel{\mathop:}=\varphi\ast W_n, where the trigonometric polynomials WnW_n are defined by (2.2). The required result is proved for each function φn\varphi_n. Hence, we have

[φ(U,V),Q]=∑n≥0[φn(U,V),Q].[\varphi(U,V),Q]=\sum_{n\geq0}[\varphi_n(U,V),Q].

and

∥[φ(U,V),Q]∥Sp≤∑n≥0∥[φn(U,V),Q]∥Sp≤const⁡∑n≥02n∥φn∥L∞(T2)(∥[U,Q]∥Sp+∥[V,Q]∥Sp)=const⁡∥φ∥B∞,11(T2)(∥[U,Q]∥Sp+∥[V,Q]∥Sp).\begin{aligned} \lVert[\varphi(U,V),Q]\rVert_{\mathcal{S}_p}\leq\sum_{n\geq0}\lVert[\varphi_n(U,V),Q]\rVert_{\mathcal{S}_p} \\ &\leq\operatorname{const}\sum_{n\geq0}2^n\lVert\varphi_n\rVert_{L^\infty(\mathbb{T}^2)}\left(\lVert[U,Q]\rVert_{\mathcal{S}_p}+\lVert[V,Q]\rVert_{\mathcal{S}_p}\right) \\ &=\operatorname{const}\lVert\varphi\rVert_{B^1_{\infty,1}(\mathbb{T}^2)}\left(\lVert[U,Q]\rVert_{\mathcal{S}_p}+\lVert[V,Q]\rVert_{\mathcal{S}_p}\right). \end{aligned}

Applying Theorem 4.1 in the case Q=ψ(U,V)Q=\psi(U,V), where ψ∈B∞,11(T2)\psi\in B^1_{\infty,1}(\mathbb{T}^2), we obtain the following theorem.

Theorem 4.2. Let UU and VV be pp-almost commuting unitary operators and let φ\varphi and ψ\psi be functions in the Besov class B∞,11(T2)B^1_{\infty,1}(\mathbb{T}^2). Then

[φ(U,V),ψ(U,V)]=∭φ(ζ,τ1)−φ(ζ,τ2)τ1−τ2 dEU(ζ) [V,ψ(U,V)] dEV(τ2)[\varphi(U,V),\psi(U,V)]=\iiint\frac{\varphi(\zeta,\tau_1)-\varphi(\zeta,\tau_2)}{\tau_1-\tau_2}\,dE_U(\zeta)\,[V,\psi(U,V)]\,dE_V(\tau_2)
+∭φ(ζ1,τ)−φ(ζ2,τ)ζ1−ζ2 dEU(ζ1) [U,ψ(U,V)] dEU(ζ2) dEV(τ).+\iiint\frac{\varphi(\zeta_1,\tau)-\varphi(\zeta_2,\tau)}{\zeta_1-\zeta_2}\,dE_U(\zeta_1)\,[U,\psi(U,V)]\,dE_U(\zeta_2)\,dE_V(\tau).

and

∥[φ(U,V),ψ(U,V)]∥Sp≤const⁡∥φ∥B∞,11(T2)∥ψ∥B∞,11(T2)∥[U,V]∥Sp\lVert[\varphi(U,V),\psi(U,V)]\rVert_{\mathcal{S}_p}\leq\operatorname{const}\lVert\varphi\rVert_{B^1_{\infty,1}(\mathbb{T}^2)}\lVert\psi\rVert_{B^1_{\infty,1}(\mathbb{T}^2)}\lVert[U,V]\rVert_{\mathcal{S}_p}

for all p∈[1,2]p\in[1,2].

An extension of the Carey–Pincus trace formula

In [10] the authors discovered the following trace formula for almost commuting unitary operators UU and VV on Hilbert space. They showed in Theorem (5.10) that for trigonometric polynomials φ\varphi and ψ\psi the following trace formula holds

trace⁡[φ(U,V),ψ(U,V)]=∬T2g(ζ,τ)(∂φ∂ζ∂ψ∂τ−∂φ∂τ∂ψ∂ζ) dζ dτ(13)\operatorname{trace}[\varphi(U,V),\psi(U,V)] = \iint_{\mathbb{T}^{2}} g(\zeta,\tau)\left(\frac{\partial\varphi}{\partial\zeta}\frac{\partial\psi}{\partial\tau}-\frac{\partial\varphi}{\partial\tau}\frac{\partial\psi}{\partial\zeta}\right)\,d\zeta\,d\tau \tag*{(13)}

where gg is the so-called principal function associated with the pair (U,V)(U,V) of almost commuting unitary operators, see [10], it is integrable on T2\mathbb{T}^{2}.

The following result extends trace formula (13) to the case when both φ\varphi and ψ\psi belong to the Besov class B∞,11(T2)B^{1}_{\infty,1}(\mathbb{T}^{2}).

Theorem 5.1. Trace formula (13) holds for arbitrary functions φ\varphi and ψ\psi in the Besov class B∞,11(T2)B^{1}_{\infty,1}(\mathbb{T}^{2}).

Proof. Put φn:=φ∗Wn\varphi_n \mathrel{\mathop:}= \varphi* W_n. As usual, the WnW_n are defined by (4). By the results of [10],

trace⁡[φn(U,V),ψn(U,V)]=∬T2g(ζ,τ)(∂φn∂ζ∂ψn∂τ−∂φn∂τ∂ψn∂ζ) dζ dτ\operatorname{trace}[\varphi_n(U,V),\psi_n(U,V)] = \iint_{\mathbb{T}^{2}} g(\zeta,\tau)\left(\frac{\partial\varphi_n}{\partial\zeta}\frac{\partial\psi_n}{\partial\tau}-\frac{\partial\varphi_n}{\partial\tau}\frac{\partial\psi_n}{\partial\zeta}\right)\,d\zeta\,d\tau

for every n≥0n \ge0. Since

φ=∑n≥0φn\varphi= \sum_{n\ge0}\varphi_n

and the series converges absolutely, this proves the theorem. ■

The case of functions of contractions

Let TT be a contraction on Hilbert space, i.e. ∥T∥≤1\lVert T\rVert\le1. Recall that each contraction TT has a unique semispectral measure ET\mathcal{E}_T defined on the σ\sigma-algebra of Borel subsets of T\mathbb{T}.

Recall that a map E\mathcal{E} defined on a σ\sigma-algebra Ω\Omega with values in the space B(H)\mathcal{B}(\mathcal{H}) of bounded linear operators on a Hilbert space H\mathcal{H} is called a semispectral measure if

0≤E(Δ)≤I,Δ∈Ω,0 \le\mathcal{E}(\Delta) \le I,\qquad\Delta\in\Omega,

and E\mathcal{E} is countably additive in the strong operator topology.

The semispectral measure ET\mathcal{E}_T of a contraction TT on H\mathcal{H} can be defined by

ET(Δ)=PHEU(Δ)∣H,Δ⊂T,\mathcal{E}_T(\Delta)=P_{\mathcal{H}}E_U(\Delta)\big|_{\mathcal{H}},\qquad\Delta\subset\mathbb{T},

where U∈B(K)U\in\mathcal{B}(\mathcal{K}) is a unitary dilation of TT and PHP_{\mathcal{H}} is the orthogonal projection on K\mathcal{K} onto H\mathcal{H}. For a function ff in the disc-algebra CA\mathcal{C}_{A} and a contraction TT the following equality holds

f(T)=∫Tf(ζ) dET(ζ).f(T)=\int_{\mathbb{T}}f(\zeta)\,d\mathcal{E}_T(\zeta).

Denote by ET\mathcal{E}_T the semi-spectral measure of TT. Let φ∈CA⊗hCA\varphi\in\mathcal{C}_{A}\otimes_{\mathrm{h}}\mathcal{C}_{A}. We define φ(T,R)\varphi(T,R) for contractions TT and RR by

φ(T,R)=∬φ(ζ,τ) dER(ζ) dET(τ):=∬φ(ζ,τ) dER(ζ)I dET(τ)\varphi(T,R)=\iint\varphi(\zeta,\tau)\,d\mathcal{E}_R(\zeta)\,d\mathcal{E}_T(\tau)\mathrel{\mathop:}= \iint\varphi(\zeta,\tau)\,d\mathcal{E}_R(\zeta)I\,d\mathcal{E}_T(\tau)

In particular, φ(T,R)\varphi(T,R) is well defined for φ∈(B∞,11)+(T2)\varphi\in(B^1_{\infty,1})_+(\mathbb{T}^2). We refer the reader to [5] for more detail. Here by (B∞,11(T2))+(B^1_{\infty,1}(\mathbb{T}^2))_+ we understand the subspace of B∞,11(T2)B^1_{\infty,1}(\mathbb{T}^2) of functions that are analytic on D2\mathbb{D}^2.

Note that the definition of double operator integrals with respect to semispectral measures is similar to the case of spectral measures. Note also that double operator integrals with respect to semispectral measures appeared first in [Pe2]. We refer the reader to [5] for more information.

Theorem 6.1. Let 1≤p≤21 \le p \le2 and let TT and RR be contractions and let QQ be a bounded linear operator such that [T,Q]∈Sp[T,Q] \in S_p and [R,Q]∈Sp[R,Q] \in S_p. Suppose that φ∈(B∞,11)+(T2)\varphi\in(B^1_{\infty,1})_+(\mathbb{T}^2). Then [φ(T,R),Q]∈Sp[\varphi(T,R),Q] \in S_p,

[φ(T,R),Q]=∭φ(ζ,τ1)−φ(ζ,τ2)τ1−τ2 dET(ζ)dER(τ1)[R,Q]dER(τ2)+∭φ(ζ1,τ)−φ(ζ2,τ)ζ1−ζ2 dET(ζ1)[T,Q]dET(ζ2)dER(τ).[\varphi(T,R),Q] = \iiint\frac{\varphi(\zeta,\tau_1)-\varphi(\zeta,\tau_2)}{\tau_1-\tau_2}\,d\mathcal{E}_T(\zeta)d\mathcal{E}_R(\tau_1)[R,Q]d\mathcal{E}_R(\tau_2) + \iiint\frac{\varphi(\zeta_1,\tau)-\varphi(\zeta_2,\tau)}{\zeta_1-\zeta_2}\,d\mathcal{E}_T(\zeta_1)[T,Q]d\mathcal{E}_T(\zeta_2)d\mathcal{E}_R(\tau).

and

∥[φ(T,R),Q]∥Sp≤const⁡∥φ∥B∞,11(T2)+(∥[T,Q]∥Sp+∥[R,Q]∥Sp).\lVert[\varphi(T,R),Q]\rVert_{S_p} \le\operatorname{const}\lVert\varphi\rVert_{B^1_{\infty,1}(\mathbb{T}^2)_+}\left(\lVert[T,Q]\rVert_{S_p}+\lVert[R,Q]\rVert_{S_p}\right).

This theorem can be proved in the same way as Theorem 4.1 but instead Theorems 5.1 and 5.2 in [1] we can apply Theorem 3.4 in [4].

The following theorem can be proved in the same way as Theorem 4.2.

Theorem 6.2. Let TT and RR be pp-almost commuting contractions and let φ\varphi and ψ\psi be functions in the Besov class (B∞,11(T2))+(B^1_{\infty,1}(\mathbb{T}^2))_+. Suppose that 1≤p≤21 \le p \le2. Then

[φ(T,R),ψ(T,R)]=∭φ(ζ,τ1)−φ(ζ,τ2)τ1−τ2 dET(ζ)dER(τ1)[V,ψ(U,V)]dER(τ2)+∭φ(ζ1,τ)−φ(ζ2,τ)ζ1−ζ2 dET(ζ1)[U,ψ(U,V)]dET(ζ2)dER(τ).[\varphi(T,R),\psi(T,R)] = \iiint\frac{\varphi(\zeta,\tau_1)-\varphi(\zeta,\tau_2)}{\tau_1-\tau_2}\,d\mathcal{E}_T(\zeta)d\mathcal{E}_R(\tau_1)[V,\psi(U,V)]d\mathcal{E}_R(\tau_2) + \iiint\frac{\varphi(\zeta_1,\tau)-\varphi(\zeta_2,\tau)}{\zeta_1-\zeta_2}\,d\mathcal{E}_T(\zeta_1)[U,\psi(U,V)]d\mathcal{E}_T(\zeta_2)d\mathcal{E}_R(\tau).

and

∥[φ(T,R),ψ(T,R)]∥Sp≤const⁡∥φ∥B∞,11(T2)∥ψ∥B∞,11(T2)∥[T,R]∥Sp.\lVert[\varphi(T,R),\psi(T,R)]\rVert_{S_p} \le\operatorname{const}\lVert\varphi\rVert_{B^1_{\infty,1}(\mathbb{T}^2)}\lVert\psi\rVert_{B^1_{\infty,1}(\mathbb{T}^2)}\lVert[T,R]\rVert_{S_p}.

The case p>2p > 2

In this section we prove that the commutator estimates for p∈[1,2]p \in[1,2] in Theorem 4.1 and Theorem 6.1 do not hold for p>2p > 2.

Theorem 7.1. Let M>0M > 0 and let 2<p≤∞2 < p \le\infty. Then there exist unitary operators UU and VV, a bounded operator QQ and an analytic polynomial ff in two variables such that

∥[f(U,V),Q]∥Sp>M∥f∥B∞,11(T2)(∥[U,Q]∥Sp+∥[V,Q]∥Sp).\lVert[f(U,V),Q]\rVert_{S_p} > M\lVert f\rVert_{B^1_{\infty,1}(\mathbb{T}^2)}\left(\lVert[U,Q]\rVert_{S_p}+\lVert[V,Q]\rVert_{S_p}\right).

To prove the theorem, we are going to use Theorem 7.3 in [5].

Theorem 7.3 in [5]. Let M>0M>0 and let 2<p≤∞2<p\le\infty. Then there exist unitary operators U1U_1, U2U_2, VV and an analytic polynomial ff in two variables such that

∥f(U1,V)−f(U2,V)∥Sp>M∥f∥B∞,11(T2)∥U1−U2∥Sp.\lVert f(U_1,V)-f(U_2,V)\rVert_{\mathcal{S}_p}>M\lVert f\rVert_{B^{1}_{\infty,1}(\mathbb{T}^2)}\lVert U_1-U_2\rVert_{\mathcal{S}_p}.

Let us introduce some notation.

Let ff be an analytic polynomial and let p∈[1,∞]p\in[1,\infty].

Denote by ∥f∥Lip⁡(Sp)\lVert f\rVert_{\operatorname{Lip}(\mathcal{S}_p)} the minimal constant C≥0C\ge0 such that

∥f(U1,V1)−f(U2,V2)∥Sp≤Cmax⁡{∥U1−U2∥Sp,∥V1−V2∥Sp}\lVert f(U_1,V_1)-f(U_2,V_2)\rVert_{\mathcal{S}_p}\le C\max\{\lVert U_1-U_2\rVert_{\mathcal{S}_p},\lVert V_1-V_2\rVert_{\mathcal{S}_p}\}

for all unitary operators U1U_1, U2U_2, V1V_1 and V2V_2.

Denote by ∥f∥CLip⁡(Sp)\lVert f\rVert_{\operatorname{CLip}(\mathcal{S}_p)} the minimal constant C≥0C\ge0 such that

∥f(U,V)Q−Qf(U,V)∥Sp≤Cmax⁡{∥UQ−QU∥Sp,∥VQ−QV∥Sp}\lVert f(U,V)Q-Qf(U,V)\rVert_{\mathcal{S}_p}\le C\max\{\lVert UQ-QU\rVert_{\mathcal{S}_p},\lVert VQ-QV\rVert_{\mathcal{S}_p}\}

for all unitary operators UU, VV and all bounded operators QQ.

We need the following lemma.

Lemma 7.2. Let 1≤p≤∞1\le p\le\infty and let ff be an analytic polynomial. Then

∥f∥Lip⁡(Sp)≤∥f∥CLip⁡(Sp).\lVert f\rVert_{\operatorname{Lip}(\mathcal{S}_p)}\le\lVert f\rVert_{\operatorname{CLip}(\mathcal{S}_p)}.

Proof. We have to prove that

∥f(U1,V1)−f(U2,V2)∥≤∥f∥CLip⁡(Sp)max⁡{∥U1−U2∥Sp,∥V1−V2∥Sp}\lVert f(U_1,V_1)-f(U_2,V_2)\rVert\le\lVert f\rVert_{\operatorname{CLip}(\mathcal{S}_p)}\max\{\lVert U_1-U_2\rVert_{\mathcal{S}_p},\lVert V_1-V_2\rVert_{\mathcal{S}_p}\}

for arbitrary unitary operators U1U_1, U2U_2, V1V_1 and V2V_2.

Consider the unitary operators UU and VV on H⊕H\mathcal{H}\oplus\mathcal{H} defined by

U=(U100U2)andV=(V100V2).U=\begin{pmatrix}U_1&0\\0&U_2\end{pmatrix} \qquad\text{and}\qquad V=\begin{pmatrix}V_1&0\\0&V_2\end{pmatrix}.

We define now the bounded operator QQ on H⊕H\mathcal{H}\oplus\mathcal{H} by

Q=(0I00).Q=\begin{pmatrix}0&I\\0&0\end{pmatrix}.

Clearly,

∥f(U,V)Q−Qf(U,V)∥Sp≤∥f∥CLip⁡(Sp)max⁡{∥UQ−QU∥Sp,∥VQ−QV∥Sp}.\lVert f(U,V)Q-Qf(U,V)\rVert_{\mathcal{S}_p}\le\lVert f\rVert_{\operatorname{CLip}(\mathcal{S}_p)}\max\{\lVert UQ-QU\rVert_{\mathcal{S}_p},\lVert VQ-QV\rVert_{\mathcal{S}_p}\}.

It remains to note that

f(U,V)Q−Qf(U,V)=(0f(U1,V1)−f(U2,V2)00).f(U,V)Q-Qf(U,V)=\begin{pmatrix}0&f(U_1,V_1)-f(U_2,V_2)\\0&0\end{pmatrix}.

and

UQ−QU=(0U1−U200)andVQ−QV=(0V1−V200).UQ-QU=\begin{pmatrix}0&U_1-U_2\\0&0\end{pmatrix} \qquad\text{and}\qquad VQ-QV=\begin{pmatrix}0&V_1-V_2\\0&0\end{pmatrix}.

Now Theorem 7.1 is an obvious consequence of Theorem 7.3 in [5].

A.B. Aleksandrov St.Petersburg State University Universitetskaya nab., 7/9 199034 St.Petersburg, Russia

St.Petersburg Department Steklov Institute of Mathematics Russian Academy of Sciences Fontanka 27, 191023 St.Petersburg Russia email: [email protected]

V.V. Peller St.Petersburg State University Universitetskaya nab., 7/9 199034 St.Petersburg, Russia

St.Petersburg Department Steklov Institute of Mathematics Russian Academy of Sciences Fontanka 27, 191023 St.Petersburg Russia email: [email protected]

Funding. The research on § 2-4 is supported by Russian Science Foundation [grant number 23-11-00153]. The research on § 5–§ 7 is supported by a grant of the Government of the Russian Federation for the state support of scientific research, carried out under the supervision of leading scientists, agreement 075-15-2025-013.

Conflict of interests. The authors declare no conflict of interests,

References

  1. [1]A.B. Aleksandrov, F.L. Nazarov and V.V. Peller, Functions of noncommuting self-adjoint operators under perturbation and estimates of triple operator integrals, Advances in Math. 295 (2016), 1-52.arxiv.org/abs/1505.07173
  2. [2]Aleksandrov A. B. , Peller V. V., Operator Lipschitz functions, Uspekhi Matem. Nauk. 71:4 (2016), 3–106 (Russian). English transl.: Russian Math. Surveys 71:4 (2016), 605–702.DOI
  3. [3]A. B. Aleksandrov and V.V. Peller, Functions of almost commuting operators and an extension of the Helton–Howe trace formula, J. Funct. Anal. 271 (2016), 3300–3322.arxiv.org/abs/1508.04702
  4. [4]A.B. Aleksandrov and V.V. Peller, Multiple operator integrals, Haagerup and Haagerup-like tensor products, and operator ideals, Bull. London Math. Soc. 49 (2017), 463–479.
  5. [5]A.B. Aleksandrov and V.V. Peller, Functions of perturbed pairs of noncommuting contractions, Izvestia RAN, 84 (2020), 538–581 (Russian). English Transl.: Izvestiya: Mathematics 84 (2020), 659–682.arxiv.org/abs/1808.08566
  6. [6]A.B. Aleksandrov and V.V. Peller, Haagerup tensor products and Schur multipliers, Algebra i Analiz 36:5 (2024), 70–85.arxiv.org/abs/2410.22523
  7. [7]M.S. Birman and M.Z. Solomyak, Double Stieltjes operator integrals, Problems of Math. Phys., Leningrad. Univ. 1 (1966), 33–67 (Russian). English transl.: Topics Math. Physics 1 (1967), 25–54, Consultants Bureau Plenum Publishing Corporation, New York.
  8. [9]M.S. Birman and M.Z. Solomyak, Double Stieltjes operator integrals. III, Problems of Math. Phys., Leningrad. Univ. 6 (1973), 27–53 (Russian).
  9. [10]R.W. Carey and J.D. Pincus, Almost commuting pairs of unitary operators and flat currents, Int. Equat and Operator Theory 4 (1981), 45–122.DOI
  10. [12]I.C. Gohberg and M.G. Krein, Introduction to the theory of linear nonselfadjoint operators in Hilbert space, Nauka, Moscow, 1965.
  11. [13]J.W. Helton and R. Howe, Integral operators, commutators, traces, index, and homology, in “Lecture Notes in Math.”, vol. 345, pp. 141–209, Springer-Verlag, New York, 1973.DOI
  12. [14]K. Juschenko, I.G. Todorov and L. Turowska, Multidimensional operator multipliers, Trans. Amer. Math. Soc. 361 (2009), 4683–4720.arxiv.org/abs/math/0701645
  13. [16]B.S. Pavlov, On multiple operator integrals, Problems of Math. Anal., No. 2: Linear Operators and Operator Equations (Russian), 99–122. Izdat. Leningrad. Univ., Leningrad, 1969.
  14. [17]J. Peetre, New thoughts on Besov spaces, Duke Univ. Press., Durham, NC, 1976.
  15. [18]V.V. Peller, Hankel operators in the theory of perturbations of unitary and self-adjoint operators, Funktsional. Anal. i Prilozhen. 19:2 (1985), 37–51 (Russian). English transl.: Funct. Anal. Appl. 19 (1985), 111–123.DOI
  16. [20]V.V. Peller, Functional calculus for a pair of almost commuting self-adjoint operators, J. Funct. Anal. 112 (1993), 325–345.
  17. [22]V.V. Peller, Multiple operator integrals in perturbation theory, Bull. Math. Sci. 6 (2016), 15–88.arxiv.org/abs/1509.02803
  18. [23]V.V. Peller, Besov spaces in operator theory, Russian Math. Surveys 79:1 (2024), 1–52.arxiv.org/abs/2402.09853

Paper details

vlnc:W_sqhn4y33
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC0-1.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents