Showing 1–12 of 12 results for author: Mihailescu, M

Convergence rates for Poisson learning to a Poisson equation with measure data

Authors: Leon Bungert, Jeff Calder, Max Mihailescu, Kodjo Houssou, Amber Yuan

Abstract: In this paper we prove discrete to continuum convergence rates for Poisson Learning, a graph-based semi-supervised learning algorithm that is based on solving the graph Poisson equation with a source term consisting of a linear combination of Dirac deltas located at labeled points and carrying label information. The corresponding continuum equation is a Poisson equation with measure data in a Eucl… ▽ More In this paper we prove discrete to continuum convergence rates for Poisson Learning, a graph-based semi-supervised learning algorithm that is based on solving the graph Poisson equation with a source term consisting of a linear combination of Dirac deltas located at labeled points and carrying label information. The corresponding continuum equation is a Poisson equation with measure data in a Euclidean domain $Ω\subset \mathbb{R}^d$. The singular nature of these equations is challenging and requires an approach with several distinct parts: (1) We prove quantitative error estimates when convolving the measure data of a Poisson equation with (approximately) radial function supported on balls. (2) We use quantitative variational techniques to prove discrete to continuum convergence rates on random geometric graphs with bandwidth $\varepsilon>0$ for bounded source terms. (3) We show how to regularize the graph Poisson equation via mollification with the graph heat kernel, and we study fine asymptotics of the heat kernel on random geometric graphs. Combining these three pillars we obtain $L^1$ convergence rates that scale, up to logarithmic factors, like $O(\varepsilon^{\frac{1}{d+2}})$ for general data distributions, and $O(\varepsilon^{\frac{2-σ}{d+4}})$ for uniformly distributed data, where $σ>0$. These rates are valid with high probability if $\varepsilon\gg\left({\log n}/{n}\right)^q$ where $n$ denotes the number of vertices of the graph and $q \approx \frac{1}{3d}$. △ Less

Submitted 9 July, 2024; originally announced July 2024.

MSC Class: 35J05; 35A35; 05C80; 35J05

On a non-homogeneous eigenvalue problem involving a potential: an Orlicz-Sobolev space setting

Authors: Mihai Mihăilescu, Vicenţiu Rădulescu, Dušan Repovš

Abstract: In this paper we study a non-homogeneous eigenvalue problem involving variable growth conditions and a potential $V$. The problem is analyzed in the context of Orlicz-Sobolev spaces. Connected with this problem we also study the optimization problem for the particular eigenvalue given by the infimum of the Rayleigh quotient associated to the problem with respect to the potential $V$ when $V$ lies… ▽ More In this paper we study a non-homogeneous eigenvalue problem involving variable growth conditions and a potential $V$. The problem is analyzed in the context of Orlicz-Sobolev spaces. Connected with this problem we also study the optimization problem for the particular eigenvalue given by the infimum of the Rayleigh quotient associated to the problem with respect to the potential $V$ when $V$ lies in a bounded, closed and convex subset of a certain variable exponent Lebesgue space. △ Less

Submitted 25 August, 2016; originally announced August 2016.

MSC Class: 35D05; 35J60; 35J70; 58E05; 68T40; 76A02

Journal ref: J. Math. Pures Appl. 93:2 (2010), 132-148

On a PDE involving the ${\cal A}_{p(\cdot)}$-Laplace operator

Authors: Mihai Mihăilescu, Dušan Repovš

Abstract: This paper establishes existence of solutions for a partial differential equation in which a differential operator involving variable exponent growth conditions is present. This operator represents a generalization of the $p(\cdot)$-Laplace operator, i.e. $Δ_{p(\cdot)}u={\rm div}(|\nabla u|^{p(\cdot)-2}\nabla u)$, where $p(\cdot)$ is a continuous function. The proof of the main result is based on… ▽ More This paper establishes existence of solutions for a partial differential equation in which a differential operator involving variable exponent growth conditions is present. This operator represents a generalization of the $p(\cdot)$-Laplace operator, i.e. $Δ_{p(\cdot)}u={\rm div}(|\nabla u|^{p(\cdot)-2}\nabla u)$, where $p(\cdot)$ is a continuous function. The proof of the main result is based on Schauder's fixed point theorem combined with adequate variational arguments. The function space setting used here makes appeal to the variable exponent Lebesgue and Sobolev spaces. △ Less

Submitted 16 March, 2016; originally announced March 2016.

MSC Class: 35B38; 35J60; 35J92

Journal ref: Nonlinear Anal. 75:2 (2012), 975-981

Multiple solutions for a nonlinear and non-homogeneous problem in Orlicz-Sobolev spaces

Authors: Mihai Mihăilescu, Dušan Repovš

Abstract: We study a non-homogeneous boundary value problem in a smooth bounded domain in $\mathbb{R}^N$. We prove the existence of at least two nonnegative and non-trivial weak solutions. Our approach relies on Orlicz-Sobolev spaces theory combined with adequate variational methods and a variant of Mountain Pass Lemma. We study a non-homogeneous boundary value problem in a smooth bounded domain in $\mathbb{R}^N$. We prove the existence of at least two nonnegative and non-trivial weak solutions. Our approach relies on Orlicz-Sobolev spaces theory combined with adequate variational methods and a variant of Mountain Pass Lemma. △ Less

Submitted 16 March, 2016; originally announced March 2016.

Comments: arXiv admin note: text overlap with arXiv:math/0607553

MSC Class: 35D30; 35J60; 35J70; 46N20; 58E05

Journal ref: Appl. Math. Comp. 217:14 (2011), 6624-6632

An eigenvalue problem involving a degenerate and singular elliptic operator

Authors: Mihai Mihăilescu, Dušan Repovš

Abstract: We study an eigenvalue problem involving a degenerate and singular elliptic operator on the whole space $\mathbb{R}^N$. We prove the existence of an unbounded and increasing sequence of eigenvalues. Our study generalizes to the case of degenerate and singular operators a result of A. Szulkin and M. Willem. We study an eigenvalue problem involving a degenerate and singular elliptic operator on the whole space $\mathbb{R}^N$. We prove the existence of an unbounded and increasing sequence of eigenvalues. Our study generalizes to the case of degenerate and singular operators a result of A. Szulkin and M. Willem. △ Less

Submitted 16 March, 2016; originally announced March 2016.

MSC Class: 35J20; 35J60; 35J70

Journal ref: Bull. Belg. Math. Soc. Simon Stevin 18:5 (2011), 839-847

Neumann problems associated to nonhomogeneous differential operators in Orlicz--Sobolev spaces

Authors: Mihai Mihailescu, Vicentiu Radulescu

Abstract: We study a nonlinear Neumann boundary value problem associated to a nonhomogeneous differential operator. Taking into account the competition between the nonlinearity and the bifurcation parameter, we establish sufficient conditions for the existence of nontrivial solutions in a related Orlicz--Sobolev space. We study a nonlinear Neumann boundary value problem associated to a nonhomogeneous differential operator. Taking into account the competition between the nonlinearity and the bifurcation parameter, we establish sufficient conditions for the existence of nontrivial solutions in a related Orlicz--Sobolev space. △ Less

Submitted 13 December, 2007; originally announced December 2007.

MSC Class: 35D05; 35J60; 35J70; 58E05; 68T40; 76A02

Journal ref: Annales de l'Institut Fourier (2008) vol. 58

A continuous spectrum for nonhomogeneous differential operators in Orlicz-Sobolev spaces

Authors: Mihai Mihailescu, Vicentiu Radulescu

Abstract: We study the nonlinear eigenvalue problem $-{\rm div}(a(|\nabla u|)\nabla u)=λ|u|^{q(x)-2}u$ in $Ω$, $u=0$ on $\partialΩ$, where $Ω$ is a bounded open set in $\RR^N$ with smooth boundary, $q$ is a continuous function, and $a$ is a nonhomogeneous potential. We establish sufficient conditions on $a$ and $q$ such that the above nonhomogeneous quasilinear problem has continuous families of eigenvalu… ▽ More We study the nonlinear eigenvalue problem $-{\rm div}(a(|\nabla u|)\nabla u)=λ|u|^{q(x)-2}u$ in $Ω$, $u=0$ on $\partialΩ$, where $Ω$ is a bounded open set in $\RR^N$ with smooth boundary, $q$ is a continuous function, and $a$ is a nonhomogeneous potential. We establish sufficient conditions on $a$ and $q$ such that the above nonhomogeneous quasilinear problem has continuous families of eigenvalues. The proofs rely on elementary variational arguments. The abstract results of this paper are illustrated by the cases $a(t)=t^{p-2}\log (1+t^r)$ and $a(t)= t^{p-2} [\log (1+t)]^{-1}$. △ Less

Submitted 6 November, 2007; originally announced November 2007.

MSC Class: 35D05; 35J60; 35J70; 58E05; 68T40; 76A02

Continuous spectrum for a class of nonhomogeneous differential operators

Authors: Mihai Mihailescu, Vicentiu Radulescu

Abstract: We study the boundary value problem $-{\rm div}((|\nabla u|^{p_1(x)-2}+|\nabla u|^{p_2(x)-2})\nabla u)=λ|u|^{q(x)-2}u$ in $Ω$, $u=0$ on $\partialΩ$, where $Ω$ is a bounded domain in $\RR^N$ with smooth boundary, $λ$ is a positive real number, and the continuous functions $p_1$, $p_2$, and $q$ satisfy $1<p_2(x)<q(x)<p_1(x)<N$ and $\max_{y\in\barΩ}q(y)<\frac{N p_2(x)}{N-p_2(x)}$ for any… ▽ More We study the boundary value problem $-{\rm div}((|\nabla u|^{p_1(x)-2}+|\nabla u|^{p_2(x)-2})\nabla u)=λ|u|^{q(x)-2}u$ in $Ω$, $u=0$ on $\partialΩ$, where $Ω$ is a bounded domain in $\RR^N$ with smooth boundary, $λ$ is a positive real number, and the continuous functions $p_1$, $p_2$, and $q$ satisfy $1<p_2(x)<q(x)<p_1(x)<N$ and $\max_{y\in\barΩ}q(y)<\frac{N p_2(x)}{N-p_2(x)}$ for any $x\in\barΩ$. The main result of this paper establishes the existence of two positive constants $λ_0$ and $λ_1$ with $λ_0\leqλ_1$ such that any $λ\in[λ_1,\infty)$ is an eigenvalue, while any $λ\in(0,λ_0)$ is not an eigenvalue of the above problem. △ Less

Submitted 27 June, 2007; originally announced June 2007.

MSC Class: 35D05; 35J60; 35J70; 58E05; 68T40; 76A02

A multiplicity result for a nonlinear degenerate problem arising in the theory of electrorheological fluids

Authors: Mihai Mihailescu, Vicentiu Radulescu

Abstract: We study a Dirichlet boundary value problem associated to an anisotropic differential operator on a smooth bounded of $\Bbb R^N$. Our main result establishes the existence of at least two different non-negative solutions, provided a certain parameter lies in a certain range. Our approach relies on the variable exponent theory of generalized Lebesgue-Sobolev spaces, combined with adequate variati… ▽ More We study a Dirichlet boundary value problem associated to an anisotropic differential operator on a smooth bounded of $\Bbb R^N$. Our main result establishes the existence of at least two different non-negative solutions, provided a certain parameter lies in a certain range. Our approach relies on the variable exponent theory of generalized Lebesgue-Sobolev spaces, combined with adequate variational methods and a variant of Mountain Pass lemma. △ Less

Submitted 21 July, 2006; originally announced July 2006.

Comments: Proceedings A of the Royal Society of London, in press

MSC Class: 35D05; 35J60; 35J70; 58E05; 68T40; 76A02

Existence and multiplicity of solutions for quasilinear nonhomogeneous problems: an Orlicz-Sobolev space setting

Authors: Mihai Mihailescu, Vicentiu Radulescu

Abstract: We study the boundary value problem $-{\rm div}(\log(1+ |\nabla u|^q)|\nabla u|^{p-2}\nabla u)=f(u)$ in $Ω$, $u=0$ on $\partialΩ$, where $Ω$ is a bounded domain in $\RR^N$ with smooth boundary. We distinguish the cases where either $f(u)=-λ|u|^{p-2}u+|u|^{r-2}u$ or $f(u)=λ|u|^{p-2}u-|u|^{r-2}u$, with $p$, $q>1$, $p+q<\min\{N,r\}$, and $r<(Np-N+p)/(N-p)$. In the first case we show the existence o… ▽ More We study the boundary value problem $-{\rm div}(\log(1+ |\nabla u|^q)|\nabla u|^{p-2}\nabla u)=f(u)$ in $Ω$, $u=0$ on $\partialΩ$, where $Ω$ is a bounded domain in $\RR^N$ with smooth boundary. We distinguish the cases where either $f(u)=-λ|u|^{p-2}u+|u|^{r-2}u$ or $f(u)=λ|u|^{p-2}u-|u|^{r-2}u$, with $p$, $q>1$, $p+q<\min\{N,r\}$, and $r<(Np-N+p)/(N-p)$. In the first case we show the existence of infinitely many weak solutions for any $λ>0$. In the second case we prove the existence of a nontrivial weak solution if $λ$ is sufficiently large. Our approach relies on adequate variational methods in Orlicz-Sobolev spaces. △ Less

Submitted 7 June, 2006; originally announced June 2006.

MSC Class: 35D05; 35J60; 35J70; 46N20

On a nonhomogeneous quasilinear eigenvalue problem in Sobolev spaces with variable exponent

Authors: Mihai Mihailescu, Vicentiu Radulescu

Abstract: We consider the nonlinear eigenvalue problem $-{\rm div}(|\nabla u|^{p(x)-2}\nabla u)=λ|u|^{q(x)-2}u$ in $Ω$, $u=0$ on $\partialΩ$, where $Ω$ is a bounded open set in $\RR^N$ with smooth boundary and $p$, $q$ are continuous functions on $\barΩ$ such that $1<\inf\_Ωq< \inf\_Ωp<\sup\_Ωq$, $\sup\_Ωp<N$, and $q(x)<Np(x)/(N-p(x))$ for all $x\in\barΩ$. The main result of this paper establishes that an… ▽ More We consider the nonlinear eigenvalue problem $-{\rm div}(|\nabla u|^{p(x)-2}\nabla u)=λ|u|^{q(x)-2}u$ in $Ω$, $u=0$ on $\partialΩ$, where $Ω$ is a bounded open set in $\RR^N$ with smooth boundary and $p$, $q$ are continuous functions on $\barΩ$ such that $1<\inf\_Ωq< \inf\_Ωp<\sup\_Ωq$, $\sup\_Ωp<N$, and $q(x)<Np(x)/(N-p(x))$ for all $x\in\barΩ$. The main result of this paper establishes that any $λ>0$ sufficiently small is an eigenvalue of the above nonhomogeneous quasilinear problem. The proof relies on simple variational arguments based on Ekeland's variational principle. △ Less

Submitted 7 June, 2006; originally announced June 2006.

MSC Class: 35D05; 35J60; 35J70; 58E05; 68T40; 76A02

Hidden classical symmetry in quantum spaces at roots of unity : From q-sphere to fuzzy sphere

Authors: Antal Jevicki, Mihail Mihailescu, Sanjaye Ramgoolam

Abstract: We study relations between different kinds of non-commutative spheres which have appeared in the context of ADS/CFT correspondences recently, emphasizing the connections between spaces that have manifest quantum group symmetry and spaces that have manifest classical symmetry. In particular we consider the quotient $SU_q(2)/U(1)$ at roots of unity, and find its relations with the fuzzy sphere wit… ▽ More We study relations between different kinds of non-commutative spheres which have appeared in the context of ADS/CFT correspondences recently, emphasizing the connections between spaces that have manifest quantum group symmetry and spaces that have manifest classical symmetry. In particular we consider the quotient $SU_q(2)/U(1)$ at roots of unity, and find its relations with the fuzzy sphere with manifest classical SU(2) symmetry. Deformation maps between classical and quantum symmetry, the $U_q(SU(2))$ module structure of quantum spheres and the structure of indecomposable representations of $U_q(SU(2))$ at roots of unity conspire in an interesting way to allow the relation between manifestly $U_q(SU(2)$ symmetric spheres and manifestly U(SU(2)) symmetric spheres. The relation suggests that a subset of field theory actions on the q-sphere are equivalent to actions on the fuzzy sphere. The results here are compatible with the proposal that quantum spheres at roots of unity appear as effective geometries which account for finite N effects in the ADS/CFT correspondence. △ Less

Submitted 10 December, 2000; v1 submitted 23 August, 2000; originally announced August 2000.

Comments: 19 pages in harvmac big, 5 figures; v2: refs added ; v3: more refs added

Report number: Brown-Het-1229