Source author record

Adam Sikora

Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.

ResearcherUnclaimed source record

Catalog footprint

Research overview

26works
5topics
4close collaborators

Published work

26 publications

preprint2021arXiv

Vertical and horizontal Square Functions on a Class of Non-Doubling Manifolds

We consider a class of non-doubling manifolds M\mathcal{M} that are the connected sum of a finite number of NN-dimensional manifolds of the form Rni×Mi\mathbb{R}^{n_{i}} \times \mathcal{M}_{i}. Following on from the work of Hassell and the second author \cite{hs2019}, a particular decomposition of the resolvent operators (Δ+k2)−M(Δ+ k^{2})^{-M}, for M∈N∗M \in \mathbb{N}^{*}, will be used to demonstrate that the vertical square function operator Sf(x):=(∫0∞∣t∇(I+t2Δ)−Mf(x)∣2dtt)12 Sf(x) := \left( \int^{\infty}_{0} \left|t \nabla (I + t^{2} Δ)^{-M}f(x)\right|{2} \frac{dt}{t}\right)^{\frac{1}{2}} is bounded on Lp(M)L^{p}(\mathcal{M}) for 1<p<nmin=min⁡ini1 < p < n_{min} = \min_{i} n_{i} and weak-type (1,1)(1,1). In addition, it will be proved that the reverse inequality ∥f∥p≲∥Sf∥p\left\Vert f \right\Vert_{p} \lesssim \left\Vert S f\right\Vert_{p} holds for p∈(nmin′,nmin)p \in (n_{min}',n_{min}) and that SS is unbounded for p≥nminp \geq n_{min} provided 2M<nmin2 M < n_{min}. Similarly, for M>1M > 1, this method of proof will also be used to ascertain that the horizontal square function operator sf(x):=(∫0∞∣t2Δ(I+t2Δ)−Mf(x)∣2 dtt)12 sf(x) := \left(\int^{\infty}_{0} \left|t^{2}Δ(I + t^{2} Δ)^{-M}f(x)\right|^{2} \, \frac{dt}{t}\right)^{\frac{1}{2}} is bounded on Lp(M)L^{p}(\mathcal{M}) for all 1<p<∞1 < p < \infty and weak-type (1,1)(1,1). Hence, for p≥nminp \geq n_{min}, the vertical and horizontal square function operators are not equivalent and their corresponding Hardy spaces HpH^p do not coincide.

preprint2020arXiv

Probabilistic approach to quantum separation effect for Feynman-Kac semigroup

Quantum tunnelling phenomenon allows a particle in Schrödinger mechanics tunnels through a barrier that it classically could not overcome. Even the infinite potentials do not always form impenetrable barriers. We discuss an answer to the following question: What is a critical magnitude of potential, which creates impenetrable barrier and for which the corresponding Schrödinger evolution system separates? In addition we describe some quantitative estimates for the separating effect in terms of cut-off potentials.

preprint2016arXiv

Spectral multipliers via resolvent type estimates on non-homogeneous metric measure spaces

We describe a simple but surprisingly effective technique of obtaining spectral multiplier results for abstract operators which satisfy the finite propagation speed property for the corresponding wave equation propagator. We show that, in this setting, spectral multipliers follow from resolvent type estimates. The most notable point of the paper is that our approach is very flexible and can be applied even if the corresponding ambient space does not satisfy the doubling condition or if the semigroup generated by an operator is not uniformly bounded. As a corollary we obtain the LpL^p spectrum independence for several second order differential operators and recover some known results. Our examples include the Laplace-Belltrami operator on manifolds with ends and Schrödinger operators with strongly subcritical potentials.

preprint2015arXiv

Spectral multipliers, Bochner-Riesz means and uniform Sobolev inequalities for elliptic operators

This paper comprises two parts. In the first, we study LpL^p to LqL^q bounds for spectral multipliers and Bochner-Riesz means with negative index in the general setting of abstract self-adjoint operators. In the second we obtain the uniform Sobolev estimates for constant coefficients higher order elliptic operators P(D)−zP(D)-z and all z∈C\[0,∞)z\in {\mathbb C}\backslash [0, \infty), which give an extension of the second order results of Kenig-Ruiz-Sogge \cite{KRS}. Next we use perturbation techniques to prove the uniform Sobolev estimates for Schrödinger operators P(D)+VP(D)+V with small integrable potentials VV. Finally we deduce spectral multiplier estimates for all these operators, including sharp Bochner-Riesz summability results.

preprint2014arXiv

Bochner-Riesz profile of anharmonic oscillator L=−d2dx2+∣x∣{\mathcal L}=-\frac{d^2}{dx^2}+|x|

We investigate spectral multipliers, Bochner-Riesz means and convergence of eigenfunction expansion corresponding to the Schrödinger operator with anharmonic potential L=−d2dx2+∣x∣{\mathcal L}=-\frac{d^2}{dx^2}+|x|. We show that the Bochner-Riesz profile of the operator L{\mathcal L} completely coincides with such profile of the harmonic oscillator H=−d2dx2+x2{\mathcal H}=-\frac{d^2}{dx^2}+x^2. It is especially surprising because the Bochner-Riesz profile for the one-dimensional standard Laplace operator is known to be essentially different and the case of operators H{\mathcal H} and L{\mathcal L} resembles more the profile of multidimensional Laplace operators. Another surprising element of the main obtained result is the fact that the proof is not based on restriction type estimates and instead entirely new perspective have to be developed to obtain the critical exponent for Bochner-Riesz means convergence.

preprint2013arXiv

A new approach to pointwise heat kernel upper bounds on doubling metric measure spaces

On doubling metric measure spaces endowed with a strongly local regular Dirichlet form, we show some characterisations of pointwise upper bounds of the heat kernel in terms of global scale-invariant inequalities that correspond respectively to the Nash inequality and to a Gagliardo-Nirenberg type inequality when the volume growth is polynomial. This yields a new proof and a generalisation of the well-known equivalence between classical heat kernel upper bounds and relative Faber-Krahn inequalities or localized Sobolev or Nash inequalities. We are able to treat more general pointwise estimates, where the heat kernel rate of decay is not necessarily governed by the volume growth. A crucial role is played by the finite propagation speed property for the associated wave equation, and our main result holds for an abstract semigroup of operators satisfying the Davies-Gaffney estimates.

preprint2013arXiv

Analysis of degenerate elliptic operators of Grušin type

We analyze degenerate, second-order, elliptic operators HH in divergence form on $L_2(\Ri^{n}\times\Ri^{m})$. We assume the coefficients are real symmetric and a1Hδ≥H≥a2Hδa_1H_δ\geq H\geq a_2H_δ for some a1,a2>0a_1,a_2>0 where Hδ=−∇x1⋅(cδ1,δ1′(x1) ∇x1)−cδ2,δ2′(x1) ∇x22  . H_δ=-{\nabla}_{x_1}\cdot(c_{δ_1, δ'_1}(x_1)\,\nabla_{x_1})-c_{δ_2, δ'_2}(x_1)\,\nabla_{x_2}^2 \;. Here $x_1\in\Ri^n$, $x_2\in\Ri^m$ and cδi,δi′c_{δ_i, δ'_i} are positive measurable functions such that cδi,δi′(x)c_{δ_i, δ'_i}(x) behaves like ∣x∣δi|x|^{δ_i} as x→0x\to0 and ∣x∣δi′|x|^{δ_i'} as x→∞x\to\infty with δ1,δ1′∈[0,1⟩δ_1,δ_1'\in[0,1\rangle and δ2,δ2′≥0δ_2,δ_2'\geq0. Our principal results state that the submarkovian semigroup St=e−tHS_t=e^{-tH} is conservative and its kernel KtK_t satisfies bounds 0≤Kt(x ;y)≤a (∣B(x ;t1/2)∣ ∣B(y ;t1/2)∣)−1/2 0\leq K_t(x\,;y)\leq a\,(|B(x\,;t^{1/2})|\,|B(y\,;t^{1/2})|)^{-1/2} where ∣B(x ;r)∣|B(x\,;r)| denotes the volume of the ball B(x ;r)B(x\,;r) centred at xx with radius rr measured with respect to the Riemannian distance associated with HH. The proofs depend on detailed subelliptic estimations on HH, a precise characterization of the Riemannian distance and the corresponding volumes and wave equation techniques which exploit the finite speed of propagation. We discuss further implications of these bounds and give explicit examples that show the kernel is not necessarily strictly positive, nor continuous.

preprint2013arXiv

Boundedness of maximal functions on non-doubling manifolds with ends

Let MM be a manifold with ends constructed in \cite{GS} and ΔΔ be the Laplace-Beltrami operator on MM. In this note, we show the weak type (1,1)(1,1) and LpL^p boundedness of the Hardy-Littlewood maximal function and of the maximal function associated with the heat semigroup $\M_Δf(x)=\sup_{t> 0} |\exp (-tΔ)f(x)| $ on Lp(M)L^p(M) for 1<p≤∞1 < p \le \infty. The significance of these results comes from the fact that MM does not satisfies the doubling condition.

preprint2013arXiv

The limitations of the Poincar{é} inequality

We examine the validity of the Poincaré inequality for degenerate, second-order, elliptic operators HH in divergence form on $L_2(\Ri^{n}\times\Ri^{m})$. We assume the coefficients are real symmetric and a1Hδ≥H≥a2Hδa_1H_δ\geq H\geq a_2H_δ for some a1,a2>0a_1,a_2>0 where HδH_δ is a generalized Grušin operator, Hδ=−∇x1 ∣x1∣(2δ1,2δ1′) ∇x1−∣x1∣(2δ2,2δ2′) ∇x22  . H_δ=-\nabla_{x_1}\,|x_1|^{(2δ_1,2δ_1')}\,\nabla_{x_1}-|x_1|^{(2δ_2,2δ_2')}\,\nabla_{x_2}^2 \;. Here $x_1\in\Ri^n$, $x_2\in\Ri^m$, δ1,δ1′∈[0,1⟩δ_1,δ_1'\in[0,1\rangle, δ2,δ2′≥0δ_2,δ_2'\geq0 and ∣x1∣(2δ,2δ′)=∣x1∣2δ|x_1|^{(2δ,2δ')}=|x_1|^{2δ} if ∣x1∣≤1|x_1|\leq 1 and ∣x1∣(2δ,2δ′)=∣x1∣2δ′|x_1|^{(2δ,2δ')}=|x_1|^{2δ'} if ∣x1∣≥1|x_1|\geq 1. \smallskip We prove that the Poincaré inequality, formulated in terms of the Riemannian geometry corresponding to HH, is valid if n≥2n\geq 2, or if n=1n=1 and δ1∨δ1′∈[0,1/2⟩δ_1\veeδ_1'\in[0,1/2\rangle but it fails if n=1n=1 and δ1∨δ1′∈[1/2,1⟩δ_1\veeδ_1'\in[1/2,1\rangle. The failure is caused by the leading term. If δ1∈[1/2,1⟩δ_1\in[1/2, 1\rangle it is an effect of the local degeneracy ∣x1∣2δ1|x_1|^{2δ_1} but if δ1∈[0,1/2⟩δ_1\in[0, 1/2\rangle and δ1′∈[1/2,1⟩δ_1'\in [1/2,1\rangle it is an effect of the growth at infinity of ∣x1∣2δ1′|x_1|^{2δ_1'}. If n=1n=1 and δ1∈[1/2,1⟩δ_1\in[1/2, 1\rangle then the semigroup SS generated by the Friedrichs' extension of HH is not ergodic. The subspaces x1≥0x_1\geq 0 and x1≤0x_1\leq 0 are SS-invariant and the Poincaré inequality is valid on each of these subspaces. If, however, n=1n=1, δ1∈[0,1/2⟩δ_1\in[0, 1/2\rangle and δ1′∈[1/2,1⟩δ_1'\in [1/2,1\rangle then the semigroup SS is ergodic but the Poincaré inequality is only valid locally. \smallskip Finally we discuss the implication of these results for the kernel of the semigroup SS.

preprint2012arXiv

Resolvent at low energy III: the spectral measure

Let M∘M^\circ be a complete noncompact manifold and gg an asymptotically conic Riemaniann metric on M∘M^\circ, in the sense that M∘M^\circ compactifies to a manifold with boundary MM in such a way that gg becomes a scattering metric on MM. Let ΔΔ be the positive Laplacian associated to gg, and P=Δ+VP = Δ+ V, where VV is a potential function obeying certain conditions. We analyze the asymptotics of the spectral measure dE(λ)=(λ/πi)(R(λ+i0)−R(λ−i0))dE(λ) = (λ/πi) \big(R(λ+i0) - R(λ- i0) \big) of P+1/2P_+^{1/2}, where R(λ)=(P−λ2)−1R(λ) = (P - λ^2)^{-1}, as λ→0λ\to 0, in a manner similar to that done previously by the second author and Vasy, and by the first two authors. The main result is that the spectral measure has a simple, `conormal-Legendrian' singularity structure on a space which is obtained from M2×[0,λ0)M^2 \times [0, λ_0) by blowing up a certain number of boundary faces. We use this to deduce results about the asymptotics of the wave solution operators cos⁡(tP+)\cos(t \sqrt{P_+}) and sin⁡(tP+)/P+\sin(t \sqrt{P_+})/\sqrt{P_+}, and the Schrödinger propagator eitPe^{itP}, as t→∞t \to \infty. In particular, we prove the analogue of Price's law for odd-dimensional asymptotically conic manifolds. This result on the spectral measure has been used in a follow-up work by the authors (arXiv:1012.3780) to prove sharp restriction and spectral multiplier theorems on asymptotically conic manifolds.

preprint2012arXiv

Restriction and spectral multiplier theorems on asymptotically conic manifolds

The classical Stein-Tomas restriction theorem is equivalent to the statement that the spectral measure dE(λ)dE(λ) of the square root of the Laplacian on $\RR^n$ is bounded from $L^p(\RR^n)$ to $L^{p'}(\RR^n)$ for 1≤p≤2(n+1)/(n+3)1 \leq p \leq 2(n+1)/(n+3), where p′p' is the conjugate exponent to pp, with operator norm scaling as λn(1/p−1/p′)−1λ^{n(1/p - 1/p') - 1}. We prove a geometric generalization in which the Laplacian on $\RR^n$ is replaced by the Laplacian, plus suitable potential, on a nontrapping asymptotically conic manifold, which is the first time such a result has been proven in the variable coefficient setting. It is closely related to, but stronger than, Sogge's discrete L2L^2 restriction theorem, which is an O(λn(1/p−1/p′)−1)O(λ^{n(1/p - 1/p') - 1}) estimate on the Lp→Lp′L^p \to L^{p'} operator norm of the spectral projection for a spectral window of fixed length. From this, we deduce spectral multiplier estimates for these operators, including Bochner-Riesz summability results, which are sharp for pp in the range above.

preprint2012arXiv

Restriction estimates, sharp spectral multipliers and endpoint estimates for Bochner-Riesz means

We consider abstract non-negative self-adjoint operators on L2(X)L^2(X) which satisfy the finite speed propagation property for the corresponding wave equation. For such operators we introduce a restriction type condition which in the case of the standard Laplace operator is equivalent to (p,2)(p,2) restriction estimate of Stein and Tomas. Next we show that in the considered abstract setting our restriction type condition implies sharp spectral multipliers and endpoint estimates for the Bochner-Riesz summability. We also observe that this restriction estimate holds for operators satisfying dispersive or Strichartz estimates. We obtain new spectral multiplier results for several second order differential operators and recover some known results. Our examples include Schrödinger operators with inverse square potentials on $\RR^n$, the harmonic oscillator, elliptic operators on compact manifolds and Schrödinger operators on asymptotically conic manifolds.

preprint2012arXiv

Sharp spectral multipliers for operators satisfying generalized Gaussian estimates

Let LL be a non-negative self adjoint operator acting on L2(X)L^2(X) where XX is a space of homogeneous type. Assume that LL generates a holomorphic semigroup e−tLe^{-tL} whose kernels pt(x,y)p_t(x,y) satisfy generalized mm-th order Gaussian estimates. In this article, we study singular and dyadically supported spectral multipliers for abstract self-adjoint operators. We show that in this setting sharp spectral multiplier results follow from Plancherel or Stein-Tomas type estimates. These results are applicable to spectral multipliers for large classes of operators including mm-th order elliptic differential operators with constant coefficients, biharmonic operators with rough potentials and Laplace type operators acting on fractals.

preprint2012arXiv

Weighted Plancherel estimates and sharp spectral multipliers for the Grushin operators

We study the Grushin operators acting on Rx′d1×Rx"d2\R^{d_1}_{x'}\times \R^{d_2}_{x"} and defined by the formula \[ L=-\sum_{\jone=1}^{d_1}\partial_{x'_\jone}^2 - (\sum_{\jone=1}^{d_1}|x'_\jone|^2) \sum_{\jtwo=1}^{d_2}\partial_{x"_\jtwo}^2. \] We obtain weighted Plancherel estimates for the considered operators. As a consequence we prove LpL^p spectral multiplier results and Bochner-Riesz summability for the Grushin operators. These multiplier results are sharp if d1≥d2d_1 \ge d_2. We discuss also an interesting phenomenon for weighted Plancherel estimates for d1<d2d_1 <d_2. The described spectral multiplier theorem is the analogue of the result for the sublaplacian on the Heisenberg group obtained by D. Müller and E.M. Stein and by W. Hebisch.

preprint2010arXiv

L1L_1-uniqueness of degenerate elliptic operators

Let ΩΩ be an open subset of $\Ri^d$ with 0∈Ω0\in Ω. Further let HΩ=−∑i,j=1d∂i cij ∂jH_Ω=-\sum^d_{i,j=1}\partial_i\,c_{ij}\,\partial_j be a second-order partial differential operator with domain Cc∞(Ω)C_c^\infty(Ω) where the coefficients cij∈Wloc1,∞(Ωˉ)c_{ij}\in W^{1,\infty}_{\rm loc}(\barΩ) are real, cij=cjic_{ij}=c_{ji} and the coefficient matrix C=(cij)C=(c_{ij}) satisfies bounds 0<C(x)≤c(∣x∣)I0<C(x)\leq c(|x|) I for all x∈Ωx\in Ω. If ∫0∞ds sd/2 e−λ μ(s)2<∞ \int^\infty_0ds\,s^{d/2}\,e^{-λ\,μ(s)^2}<\infty for some λ>0λ>0 where μ(s)=∫0sdt c(t)−1/2μ(s)=\int^s_0dt\,c(t)^{-1/2} then we establish that HΩH_Ω is L1L_1-unique, i.e.\ it has a unique L1L_1-extension which generates a continuous semigroup, if and only if it is Markov unique, i.e.\ it has a unique L2L_2-extension which generates a submarkovian semigroup. Moreover these uniqueness conditions are equivalent with the capacity of the boundary of ΩΩ, measured with respect to HΩH_Ω, being zero. We also demonstrate that the capacity depends on two gross features, the Hausdorff dimension of subsets AA of the boundary the set and the order of degeneracy of HΩH_Ω at AA.

preprint2010arXiv

Weighted norm inequalities, Gaussian bounds and sharp spectral multipliers

Let LL be a non-negative self adjoint operator acting on L2(X)L^2(X) where XX is a space of homogeneous type. Assume that LL generates a holomorphic semigroup e−tLe^{-tL} whose kernels pt(x,y)p_t(x,y) have Gaussian upper bounds but possess no regularity in variables xx and yy. In this article, we study weighted LpL^p-norm inequalities for spectral multipliers of LL. We show sharp weighted Hörmander-type spectral multiplier theorems follow from Gaussian heat kernel bounds and appropriate L2L^2 estimates of the kernels of the spectral multipliers. These results are applicable to spectral multipliers for large classes of operators including Laplace operators acting on Lie groups of polynomial growth or irregular non-doubling domains of Euclidean spaces, elliptic operators on compact manifolds and Schrödinger operators with non-negative potentials on complete Riemannian manifolds.

preprint2009arXiv

Degenerate elliptic operators in one dimension

Let HH be the symmetric second-order differential operator on $L_2(\Ri)$ with domain $C_c^\infty(\Ri)$ and action Hφ=−(cφ′)′Hφ=-(c φ')' where $ c\in W^{1,2}_{\rm loc}(\Ri)$ is a real function which is strictly positive on $\Ri\backslash\{0\}$ but with c(0)=0c(0)=0. We give a complete characterization of the self-adjoint extensions and the submarkovian extensions of HH. In particular if ν=ν+∨ν−ν=ν_+\veeν_- where ν±(x)=±∫±x±1c−1ν_\pm(x)=\pm\int^{\pm 1}_{\pm x} c^{-1} then HH has a unique self-adjoint extension if and only if ν∉L2(0,1)ν\not\in L_2(0,1) and a unique submarkovian extension if and only if ν∉L∞(0,1)ν\not\in L_\infty(0,1). In both cases the corresponding semigroup leaves L2(0,∞)L_2(0,\infty) and L2(−∞,0)L_2(-\infty,0) invariant. In addition we prove that for a general non-negative $ c\in W^{1,\infty}_{\rm loc}(\Ri)$ the corresponding operator HH has a unique submarkovian extension.

preprint2009arXiv

Markov uniqueness of degenerate elliptic operators

Let ΩΩ be an open subset of $\Ri^d$ and HΩ=−∑i,j=1d∂icij∂jH_Ω=-\sum^d_{i,j=1}\partial_i c_{ij} \partial_j a second-order partial differential operator on L2(Ω)L_2(Ω) with domain Cc∞(Ω)C_c^\infty(Ω) where the coefficients cij∈W1,∞(Ω)c_{ij}\in W^{1,\infty}(Ω) are real symmetric and C=(cij)C=(c_{ij}) is a strictly positive-definite matrix over ΩΩ. In particular, HΩH_Ω is locally strongly elliptic. We analyze the submarkovian extensions of HΩH_Ω, i.e. the self-adjoint extensions which generate submarkovian semigroups. Our main result establishes that HΩH_Ω is Markov unique, i.e. it has a unique submarkovian extension, if and only if $\capp_Ω(\partialΩ)=0$ where $\capp_Ω(\partialΩ)$ is the capacity of the boundary of ΩΩ measured with respect to HΩH_Ω. The second main result establishes that Markov uniqueness of HΩH_Ω is equivalent to the semigroup generated by the Friedrichs extension of HΩH_Ω being conservative.

preprint2006arXiv

Analysis of degenerate elliptic operators of Grushin type

We analyze degenerate, second-order, elliptic operators HH in divergence form on L2(Rn×Rm)L_2({\bf R}^{n}\times{\bf R}^{m}). We assume the coefficients are real symmetric and a1Hδ≥H≥a2Hδa_1H_δ\geq H\geq a_2H_δ for some a1,a2>0a_1,a_2>0 where Hδ=−∇x1cδ1,δ1′(x1)∇x1−cδ2,δ2′(x1)∇x22. H_δ=-\nabla_{x_1} c_{δ_1, δ'_1}(x_1) \nabla_{x_1}-c_{δ_2, δ'_2}(x_1) \nabla_{x_2}^2 . Here x1∈Rnx_1\in{\bf R}^n, x2∈Rmx_2\in{\bf R}^m and cδi,δi′c_{δ_i, δ'_i} are positive measurable functions such that cδi,δi′(x)c_{δ_i, δ'_i}(x) behaves like ∣x∣δi|x|^{δ_i} as x→0x\to0 and ∣x∣δi′|x|^{δ_i'} as x→∞x\to\infty with δ1,δ1′∈[0,1>δ_1,δ_1'\in[0,1> and δ2,δ2′≥0δ_2,δ_2'\geq0. Our principal results state that the submarkovian semigroup St=e−tHS_t=e^{-tH} is conservative and its kernel KtK_t satisfies bounds 0≤Kt(x;y)≤a(∣B(x;t1/2)∣∣B(y;t1/2)∣)−1/2 0\leq K_t(x ;y)\leq a (|B(x ;t^{1/2})| |B(y ;t^{1/2})|)^{-1/2} where ∣B(x;r)∣|B(x ;r)| denotes the volume of the ball B(x;r)B(x ;r) centred at xx with radius rr measured with respect to the Riemannian distance associated with HH. The proofs depend on detailed subelliptic estimations on HH, a precise characterization of the Riemannian distance and the corresponding volumes and wave equation techniques which exploit the finite speed of propagation. We discuss further implications of these bounds and give explicit examples that show the kernel is not necessarily strictly positive, nor continuous.

preprint2006arXiv

Degenerate elliptic operators: capacity, flux and separation

Let S={St}t≥0S=\{S_t\}_{t\geq0} be the semigroup generated on $L_2(\Ri^d)$ by a self-adjoint, second-order, divergence-form, elliptic operator HH with Lipschitz continuous coefficients. Further let ΩΩ be an open subset of $\Ri^d$ with Lipschitz continuous boundary ∂Ω\partialΩ. We prove that SS leaves L2(Ω)L_2(Ω) invariant if, and only if, the capacity of the boundary with respect to HH is zero or if, and only if, the energy flux across the boundary is zero. The global result is based on an analogous local result.

preprint2006arXiv

Dirichlet forms and degenerate elliptic operators

It is shown that the theory of real symmetric second-order elliptic operators in divergence form on $\Ri^d$ can be formulated in terms of a regular strongly local Dirichlet form irregardless of the order of degeneracy. The behaviour of the corresponding evolution semigroup StS_t can be described in terms of a function (A,B)↦d(A;B)∈[0,∞](A,B) \mapsto d(A ;B)\in[0,\infty] over pairs of measurable subsets of $\Ri^d$. Then ∣(ϕA,StϕB)∣≤e−d(A;B)2(4t)−1∥ϕA∥2∥ϕB∥2 |(ϕ_A,S_tϕ_B)|\leq e^{-d(A;B)^2(4t)^{-1}}\|ϕ_A\|_2\|ϕ_B\|_2 for all t>0t>0 and all ϕA∈L2(A)ϕ_A\in L_2(A), ϕB∈L2(B)ϕ_B\in L_2(B). Moreover StL2(A)⊆L2(A)S_tL_2(A)\subseteq L_2(A) for all t>0t>0 if and only if d(A;Ac)=∞d(A ;A^c)=\infty where AcA^c denotes the complement of AA.

preprint2006arXiv

Second-order operators with degenerate coefficients

We consider properties of second-order operators H=−∑i,j=1d∂i cij ∂jH = -\sum^d_{i,j=1} \partial_i \, c_{ij} \, \partial_j on $\Ri^d$ with bounded real symmetric measurable coefficients. We assume that C=(cij)≥0C = (c_{ij}) \geq 0 almost everywhere, but allow for the possibility that CC is singular. We associate with HH a canonical self-adjoint viscosity operator H0H_0 and examine properties of the viscosity semigroup S(0)S^{(0)} generated by H0H_0. The semigroup extends to a positive contraction semigroup on the LpL_p-spaces with p∈[1,∞]p \in [1,\infty]. We establish that it conserves probability, satisfies L2L_2~off-diagonal bounds and that the wave equation associated with H0H_0 has finite speed of propagation. Nevertheless S(0)S^{(0)} is not always strictly positive because separation of the system can occur even for subelliptic operators. This demonstrates that subelliptic semigroups are not ergodic in general and their kernels are neither strictly positive nor Hölder continuous. In particular one can construct examples for which both upper and lower Gaussian bounds fail even with coefficients in $C^{2-\varepsilon}(\Ri^d)$ with ε>0\varepsilon > 0.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.