Nash Type Inequalities for Fractional Powers of Non-Negative Self-adjoint Operators. ( Wroclaw 2006) P.Maheux (Orléans. France)

Similar documents
Logarithmic Sobolev Inequalities

Heat Flows, Geometric and Functional Inequalities

Holomorphic functions which preserve holomorphic semigroups

Brownian Motion and Stochastic Calculus

NASH TYPE INEQUALITIES FOR FRACTIONAL POWERS OF NON-NEGATIVE SELF-ADJOINT OPERATORS

PROBABILITY: LIMIT THEOREMS II, SPRING HOMEWORK PROBLEMS

{σ x >t}p x. (σ x >t)=e at.

Functional Inequalities for Gaussian and Log-Concave Probability Measures

Intertwinings for Markov processes

Fonctions on bounded variations in Hilbert spaces

Let R be the line parameterized by x. Let f be a complex function on R that is integrable. The Fourier transform ˆf = F f is. e ikx f(x) dx. (1.

P(E t, Ω)dt, (2) 4t has an advantage with respect. to the compactly supported mollifiers, i.e., the function W (t)f satisfies a semigroup law:

PROBABILITY: LIMIT THEOREMS II, SPRING HOMEWORK PROBLEMS

The Heat Equation John K. Hunter February 15, The heat equation on a circle

ON PARABOLIC HARNACK INEQUALITY

Wiener Measure and Brownian Motion

ON THE SHAPE OF THE GROUND STATE EIGENFUNCTION FOR STABLE PROCESSES

Lebesgue s Differentiation Theorem via Maximal Functions

Math 421, Homework #9 Solutions

Notes on uniform convergence

ON MEHLER S FORMULA. Giovanni Peccati (Luxembourg University) Conférence Géométrie Stochastique Nantes April 7, 2016

Kolmogorov Equations and Markov Processes

Fine scales of decay rates of operator semigroups

1 Independent increments

Part II Probability and Measure

Probability and Measure

Kernel Density Estimation

Around Nash inequalities

Statistics, Data Analysis, and Simulation SS 2013

NEW FUNCTIONAL INEQUALITIES

On lower bounds for C 0 -semigroups

Heat kernels of some Schrödinger operators

Midterm 1 Solutions Thursday, February 26

Elliptic Operators with Unbounded Coefficients

Feller Processes and Semigroups

Heat kernels on metric measure spaces

Recall that if X is a compact metric space, C(X), the space of continuous (real-valued) functions on X, is a Banach space with the norm

Martingale Problems. Abhay G. Bhatt Theoretical Statistics and Mathematics Unit Indian Statistical Institute, Delhi

THEOREM OF OSELEDETS. We recall some basic facts and terminology relative to linear cocycles and the multiplicative ergodic theorem of Oseledets [1].

Bochner s Theorem on the Fourier Transform on R

MATH MEASURE THEORY AND FOURIER ANALYSIS. Contents

SCHWARTZ SPACES ASSOCIATED WITH SOME NON-DIFFERENTIAL CONVOLUTION OPERATORS ON HOMOGENEOUS GROUPS

Analysis Comprehensive Exam Questions Fall 2008

On semilinear elliptic equations with measure data

A Lévy-Fokker-Planck equation: entropies and convergence to equilibrium

Function spaces and multiplier operators

ELEMENTS OF PROBABILITY THEORY

Hardy-Stein identity and Square functions

d(x n, x) d(x n, x nk ) + d(x nk, x) where we chose any fixed k > N

Errata Applied Analysis

p 1 ( Y p dp) 1/p ( X p dp) 1 1 p

Banach Spaces, HW Problems & Solutions from Hunter

Examples of Dual Spaces from Measure Theory

Definition: Lévy Process. Lectures on Lévy Processes and Stochastic Calculus, Braunschweig, Lecture 2: Lévy Processes. Theorem

A Sharpened Hausdorff-Young Inequality

THEOREMS, ETC., FOR MATH 516

Simultaneous Gaussian quadrature for Angelesco systems

Stability results for Logarithmic Sobolev inequality

Bernstein s inequality and Nikolsky s inequality for R d

A Note on the Central Limit Theorem for a Class of Linear Systems 1

Concentration inequalities for non-lipschitz functions

Applications of the time derivative of the L 2 -Wasserstein distance and the free entropy dissipation

Notation. General. Notation Description See. Sets, Functions, and Spaces. a b & a b The minimum and the maximum of a and b

Rigorous Functional Integration with Applications to Nelson s and the Pauli-Fierz Model

Name: Partners: PreCalculus. Review 5 Version A

MAT 578 FUNCTIONAL ANALYSIS EXERCISES

CHAPTER 6. Differentiation

A note on the convex infimum convolution inequality

Exponential functionals of Lévy processes

THEOREMS, ETC., FOR MATH 515

MS 2001: Test 1 B Solutions

Potential theory of subordinate killed Brownian motions

Spectral theory for linear operators on L 1 or C(K) spaces

Math The Laplacian. 1 Green s Identities, Fundamental Solution

Metric Spaces Math 413 Honors Project

COMPETITIVE OR WEAK COOPERATIVE STOCHASTIC LOTKA-VOLTERRA SYSTEMS CONDITIONED ON NON-EXTINCTION. Université de Toulouse

Partial Differential Equations

Regularizations of Singular Integral Operators (joint work with C. Liaw)

Metric Spaces Math 413 Honors Project

CMS winter meeting 2008, Ottawa. The heat kernel on connected sums

Green s Functions and Distributions

6.2 Fubini s Theorem. (µ ν)(c) = f C (x) dµ(x). (6.2) Proof. Note that (X Y, A B, µ ν) must be σ-finite as well, so that.

Convergence to equilibrium of Markov processes (eventually piecewise deterministic)

Problem Set 5: Solutions Math 201A: Fall 2016

be the set of complex valued 2π-periodic functions f on R such that

Differentiation of Measures and Functions

Harnack Inequalities and Applications for Stochastic Equations

M. Ledoux Université de Toulouse, France

curvature, mixing, and entropic interpolation Simons Feb-2016 and CSE 599s Lecture 13

Reminder Notes for the Course on Distribution Theory

Strong uniqueness for stochastic evolution equations with possibly unbounded measurable drift term

1 Math 241A-B Homework Problem List for F2015 and W2016

Spectral theorems for bounded self-adjoint operators on a Hilbert space

ERRATA TO REAL ANALYSIS, 2nd edition (6th and later printings) G. B. Folland

An idea how to solve some of the problems. diverges the same must hold for the original series. T 1 p T 1 p + 1 p 1 = 1. dt = lim

Real Analysis Qualifying Exam May 14th 2016

Singular Integrals. 1 Calderon-Zygmund decomposition

Statistical inference on Lévy processes

1 Review of di erential calculus

Transcription:

Nash Type Inequalities for Fractional Powers of Non-Negative Self-adjoint Operators ( Wroclaw 006) P.Maheux (Orléans. France) joint work with A.Bendikov. European Network (HARP) (to appear in T.A.M.S) 1

Classical example: Heat semigroup on n R with dx the Lebesgue measure: T t f(x) = h t f(x) = h t (x, y) = 1 (4πt) n/ exp R n h t(x, y)f(y) dy ( x y 4t ) Af = f, f C 0 (Rn ) D T t 1 := sup{ T t f, f 1 = 1} (Ult) T t 1 = h t (0, 0) = 1 (4πt) n/, t > 0 (Sob) f n/n c( f, f), (Nash) f +4/n c( f, f) f 4/n 1,

Let (T t ) t>0 be a symmetric submarkovian semigroup acting on L (X, µ) with µ a σ-finite measure on X and let A be its generator with domain D. f D T t+s = T t T s t, s > 0 lim T tf = f (in L ) t 0 + 0 f 1 0 T t f 1 t > 0 T t f f Af = lim t 0 + t L T t = e ta acts on L p (1 p + ) space as a contraction semigroup: T t f p f p, t > 0 3

Theorem[V,C-K-S] Let ν >. The following conditions are equivalent : (Sob) f ν/ν c (Af, f), f D (Nash) f +4/ν c (Af, f) f 4/ν 1, f D L 1 (Ult) T t 1 c t ν/, t > 0 4

Fact: Nash inequality for A implies Nash inequality for all α (0, 1) (by subordination): f +4α/ν c (A α f, f) f 4α/ν 1 Nash-type (or generalized) inequality(def) Let B : [0, + [ [0, + [ be a non-decreasing function. A satisfies a Nash-type inequality if f B( f ) (Af, f), f D f 1 1 Classical case in R d (Nash) c f ( f ) /n (Af, f) with B(x) = cx /n 5

Theorem [T.Coulhon. J.F.A 141] (Ult) T t 1 m(t), t > 0. implies Nash type inequality f B( f ) (Af, f), f D, f 1 1 with B(x) = sup s>0 [ s log x sg(s)] G(s) = log m(1/s) 6

Classical example: Polynomial decay: Under the general assumptions above on the semigroup: T t 1 c =: m(t) tν/ then B(x) = cx /ν Another example: One exponential decay Proposition Let γ > 0. Assume that T t 1 c exp(1/t γ ) t > 0 Then for all f D with f 1 1, f Example : On T, [ log+ ( f )] 1+1/γ (Af, f) A = + k=1 k 1 γ k

7 Another example: Double exponential decay Proposition Let β > 0. Assume that T t 1 c e e1/tβ t > 0 Then for all f D with f 1 1, f log + ( f ) [ log+ ( f )] 1/β (Af, f) Example : On T, A = + k=1 (ln k) γ k 8

An example of Nash-type inequality without Ultracontractivity: The Ornstein-Uhlenbeck operator A = +x on L n (R, γ n ) with the Gaussian measure: γ n (x) = 1 ( ) x exp dx (π) n/ Proposition f log + ( f ) (Af, f), f 1 1 then Θ(x) = x log + (x) But (T t ) is not ultracontractive! Proof. Deduce from Log-Sobolev inequality of Gross (but slightly weaker). 9

Spectral Theory Let A = A be a non negative self-adjoint operator (unbounded) on L (X, µ) with µ a positive σ-finite Spectral theory: A = + 0 λ de λ (E λ spectral measure). Functional calculus: Let f : [0, + [ R be a Borel function, Examples: f(a) = + 0 f(λ) de λ (t > 0) f t (λ) = e tλ : f(λ) = λ α, 0 < α < + : T t = e ta semi-group A α fractional power f(λ) = log(1 + λ) : log(i + A) 10

General Question: Assume a Nash type inequality holds for A: f B( f ) (Af, f), f D, f 1 1 Is it true that g(a) satisfies a Nash type inequality for some function g 0? If yes: describe the function B g s.t: f B g( f ) (g(a)f, f) f D g, f 1 1 In particular with: g(x) = x α, 0 < α < 1 : Fractional powers g(x) = log(1 + x) : log -semigroup 11

Why are we so interested by g(a)? If g : [0, + [ [ is real and g 0 then g(a) is also non negative self adjoint operator (defined by spectral theory). Then, we can define the associated semigroup T g t = e tg(a) 1

Locally Compact Abelian Groups Setting Let G be a l.c.a group with dual group Ĝ. ( g is Bernstein function) g is a Bernstein function if g :]0, + [ [0, + [, g is C and ( 1) p f (p) 0 p N {0}. Examples: g(x) = 1 e sx, s > 0 : Poisson sgr. with jump s g(x) = x α, 0 < α < 1 : Fractional powers g(x) = log(1 + x) : Γ-semigroup Why are we so interested by g(a)? 1a

Let µ t be a convolution semigroup on G given by ˆµ t (y) = e tψ(y), y Ĝ with ψ the associated continuous negative definite function. Then (goψ) is a continuous negative definite function ( µ g t )ˆ(y) = e t (goψ)(y), y Ĝ is also a convolution semigroup! Representation formula for Berstein function g(x) = a + bx + + 0 (1 e xs ) dµ(s) with µ a positive measure on [0, + [ 1 + s dµ(s) < dµ(s) < 0 1 1b

Examples: g(x) = x α (0 < α < 1) then g(a) = A α is the fractional operator. g(x) = log(1 + x) i.e g(a) = log(i + A). The operator g(a) generates a Markov semigroup. Is it true that g(a) satisfies a Nash type inequality? The answer is yes with A α. Open problem: Nash inequality for log(i + A) (Partial results in R n and also in the abstract setting) 13

The Assumptions: Let (X, µ) be a measure space with σ-finite measure µ. Let A be a non-negative self-adjoint operator with domain D(A) L (X, µ). Suppose that the semigroup T t = e ta acts as a contraction on L 1 (X, µ) Let B : [0, + [ [0, + [ be a non-decreasing function which tends to infinity at infinity. 14

Theorem(A.Bendikov,P.M) If the operator A satisfies f B ( f ) (Af, f) f D(A), f 1 1 Then, for any α > 0, f [ B ( f )] α (A α f, f) f D(A α ), f 1 1. Picture : Operators: A A α Functions: > B(x) > B α (x) 15

16