in Bounded Domains Ariane Trescases CMLA, ENS Cachan

Similar documents
Fourier Law and Non-Isothermal Boundary in the Boltzmann Theory

YAN GUO, JUHI JANG, AND NING JIANG

ALMOST EXPONENTIAL DECAY NEAR MAXWELLIAN

arxiv: v1 [math.ap] 28 Apr 2009

Author(s) Huang, Feimin; Matsumura, Akitaka; Citation Osaka Journal of Mathematics. 41(1)

Lecture No 1 Introduction to Diffusion equations The heat equat

hal , version 1-22 Nov 2009

On the Interior Boundary-Value Problem for the Stationary Povzner Equation with Hard and Soft Interactions

Exponential tail behavior for solutions to the homogeneous Boltzmann equation

Existence and uniqueness of solutions for a diffusion model of host parasite dynamics

Some asymptotic properties of solutions for Burgers equation in L p (R)

The 2D Magnetohydrodynamic Equations with Partial Dissipation. Oklahoma State University

Kinetic models of Maxwell type. A brief history Part I

On Weak Solutions to the Linear Boltzmann Equation with Inelastic Coulomb Collisions

From Newton s law to the linear Boltzmann equation without cut-off

Piecewise Smooth Solutions to the Burgers-Hilbert Equation

Presenter: Noriyoshi Fukaya

Alignment processes on the sphere

Global well-posedness of the primitive equations of oceanic and atmospheric dynamics

New Identities for Weak KAM Theory

LMI Methods in Optimal and Robust Control

VALIDITY OF THE BOLTZMANN EQUATION

Nonlinear Analysis 71 (2009) Contents lists available at ScienceDirect. Nonlinear Analysis. journal homepage:

LARGE TIME BEHAVIOR OF THE RELATIVISTIC VLASOV MAXWELL SYSTEM IN LOW SPACE DIMENSION

Mathematical modelling of collective behavior

On continuous time contract theory

On the Boltzmann equation: global solutions in one spatial dimension

Anomalous transport of particles in Plasma physics

Exercises. T 2T. e ita φ(t)dt.

Régularité des équations de Hamilton-Jacobi du premier ordre et applications aux jeux à champ moyen

Mathematical Methods for Neurosciences. ENS - Master MVA Paris 6 - Master Maths-Bio ( )

Entropic structure of the Landau equation. Coulomb interaction

Uniformly accurate averaging numerical schemes for oscillatory evolution equations

Nonlinear representation, backward SDEs, and application to the Principal-Agent problem

Lecture No 2 Degenerate Diffusion Free boundary problems

Low Froude Number Limit of the Rotating Shallow Water and Euler Equations

A posteriori analysis of a discontinuous Galerkin scheme for a diffuse interface model

Fractal Conservation Laws: Global Smooth Solutions and Vanishing Regularization

From the Newton equation to the wave equation in some simple cases

Global existence of smooth solutions to two-dimensional compressible isentropic Euler equations for Chaplygin gases

1 Assignment 1: Nonlinear dynamics (due September

GLOBAL EXISTENCE AND ENERGY DECAY OF SOLUTIONS TO A PETROVSKY EQUATION WITH GENERAL NONLINEAR DISSIPATION AND SOURCE TERM

Short-time expansions for close-to-the-money options under a Lévy jump model with stochastic volatility

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

Hypocoercivity for kinetic equations with linear relaxation terms

Travelling waves. Chapter 8. 1 Introduction

Entropy-dissipation methods I: Fokker-Planck equations

Boltzmann asymptotics with diffuse reflection boundary conditions.

I forgot to mention last time: in the Ito formula for two standard processes, putting

From Newton s law to the linear Boltzmann equation without cut-off

MATH 425, FINAL EXAM SOLUTIONS

b i (µ, x, s) ei ϕ(x) µ s (dx) ds (2) i=1

Stability of Stochastic Differential Equations

A quantum heat equation 5th Spring School on Evolution Equations, TU Berlin

NONLINEAR PROPAGATION OF WAVE PACKETS. Ritsumeikan University, and 22

Modeling uncertainties with kinetic equations

Pointwise convergence rate for nonlinear conservation. Eitan Tadmor and Tao Tang

MATH 220 solution to homework 5

THE VLASOV-POISSON-BOLTZMANN SYSTEM FOR SOFT POTENTIALS

Stochastic optimal control with rough paths

Half of Final Exam Name: Practice Problems October 28, 2014

An asymptotic-preserving micro-macro scheme for Vlasov-BGK-like equations in the diffusion scaling

Lecture 4: Numerical solution of ordinary differential equations

Nonlinear instability of periodic BGK waves for Vlasov-Poisson system

Global regularity of a modified Navier-Stokes equation

{ } is an asymptotic sequence.

GLOBAL WELL-POSEDNESS IN SPATIALLY CRITICAL BESOV SPACE FOR THE BOLTZMANN EQUATION

Une approche hypocoercive L 2 pour l équation de Vlasov-Fokker-Planck

The Schrödinger equation with spatial white noise potential

10 Transfer Matrix Models

A stochastic particle system for the Burgers equation.

Hydrodynamic Limits for the Boltzmann Equation

Radon measure solutions for scalar. conservation laws. A. Terracina. A.Terracina La Sapienza, Università di Roma 06/09/2017

2 Formal derivation of the Shockley-Read-Hall model

Functional Analysis Exercise Class

On Asymptotic Variational Wave Equations

Exponential moments for the homogeneous Kac equation

Convergence of Finite Volumes schemes for an elliptic-hyperbolic system with boundary conditions

Determination of thin elastic inclusions from boundary measurements.

Lecture 10: Singular Perturbations and Averaging 1

THE UNIVERSITY OF WESTERN ONTARIO. Applied Mathematics 375a Instructor: Matt Davison. Final Examination December 14, :00 12:00 a.m.

Effective dynamics for the (overdamped) Langevin equation

ENERGY CONSERVATIVE SOLUTIONS TO A NONLINEAR WAVE SYSTEM OF NEMATIC LIQUID CRYSTALS

Unbiased Risk Estimation as Parameter Choice Rule for Filter-based Regularization Methods

Incoming and disappearaing solutions of Maxwell s equations. Université Bordeaux I

Contractive metrics for scalar conservation laws

analysis for transport equations and applications

Midterm Exam, Thursday, October 27

REMARKS ON THE ACOUSTIC LIMIT FOR THE BOLTZMANN EQUATION

Stationary mean-field games Diogo A. Gomes

Global conservative solutions of the Camassa-Holm equation

at time t, in dimension d. The index i varies in a countable set I. We call configuration the family, denoted generically by Φ: U (x i (t) x j (t))

Stochastic Volatility and Correction to the Heat Equation

The inviscid limit to a contact discontinuity for the compressible Navier-Stokes-Fourier system using the relative entropy method

BLOWUP THEORY FOR THE CRITICAL NONLINEAR SCHRÖDINGER EQUATIONS REVISITED

Integrodifferential Hyperbolic Equations and its Application for 2-D Rotational Fluid Flows

OPTIMAL CONVERGENCE RATES FOR THE COMPRESSIBLE NAVIER-STOKES EQUATIONS WITH POTENTIAL FORCES

Laplace s Equation. Chapter Mean Value Formulas

arxiv: v1 [math.ap] 20 Nov 2007

Homogenization and error estimates of free boundary velocities in periodic media

Transcription:

CMLA, ENS Cachan Joint work with Yan GUO, Chanwoo KIM and Daniela TONON International Conference on Nonlinear Analysis: Boundary Phenomena for Evolutionnary PDE Academia Sinica December 21, 214

Outline The system Results Convex domains Non-

The system Results The system Results Convex domains Non-

The Boltzmann equation F (t, x, v): density distribution of gas particles at time t, at point x of space Ω R 3 and with velocity v R 3. It satisfies tf + v }{{ xf = Q(F, F ), }}{{} free transport collisions The system Results Q(F 1, F 2) = Q gain (F 1, F 2) Q loss (F 1, F 2) = v u κ {F 1(u )F 2(v ) F 1(u)F 2(v)}q (θ) dω du, R 3 S 2 where v = v + [(u v) ω] ω, where u = u [(u v) ω] ω. Assume κ 1 (hard potential), q C v u ω (angular cutoff). v u

Boundary conditions Ω is a bounded domain of R 3. We define in the phase space the outgoing/incoming/grazing boundary γ + = {(x, v) Ω R 3 : v n(x) > }, (x,v) γ + (x,v) γ Ω The system Results γ = {(x, v) Ω R 3 : v n(x) < }, γ = {(x, v) Ω R 3 : v n(x) = }. n(x) x (x,v) γ - Diffuse boundary condition: writing µ(v) = exp{ v 2 /2}, (x, v) γ, F (t, x, v) = c µ µ(v) F (t, x, u) u n(x) du. u n(x)>

The system: summary Define F (t, x, v) = µ(v) f (t, x, v). We rewrite the Cauchy problem: tf + v xf + ν[ µf ]f = Γ(f, f ), f t= = f, (1) f γ = c µ µ(v) µ(u)f (u)u n(x) du, u n(x)> The system Results where ν[ µf ] = Q loss ( µf, 1), Γ(f, f ) = Q gain( µf, µf ) µ(v).

Known results for weak solutions in general bounded domains: existence theory (H ): F = µf and e θ v 2 f < ( < θ < 1/4). Theorem (Existence/uniqueness, [Guo 1], [GKTT 13]) (H ). There exists a unique solution f of (1) on [, T ) with T = T (f ). Furthermore for some < θ < θ, T < T The system Results sup e θ v 2 f (t) T e θ v 2 f. t T If e θ v 2 {f µ} 1, then T =.

Known results for weak solutions in general bounded domains: regularity In : continuity away from the grazing boundary γ [Guo 1] In non-: discontinuity created on the grazing boundary and propagated along the (grazing) characteristics [Kim 11] The system Results Continuity Ω Continuity Ω Discontinuity

New results: regularity in [GKTT 13] Assume Ω is a smooth strictly convex domain. Assume (H ) and f satisfies the diffuse boundary condition. Theorem (, 1 < p < 2) Assume p (1, 2) and f L p (Ω R 3 ). Then T T < T, sup f p p(t) + f p v.n ds dv dt t T Ω R 3 The system Results with P a polynomial. T f p p + P( e θ v 2 f ), Note: we use the notation tf = v xf ν[ µf ]f + Γ(f, f ).

New results: regularity in [GKTT 13] Assume Ω is a smooth strictly convex domain. Assume (H ) and f satisfies the diffuse boundary condition. Theorem (Weighted, p [2, ]) Assume d 2 p f L p (Ω R 3 ). Then T < T, if p [2, [ T sup dp 2 f p p(t) + dp 2 f p v.n ds dv dt t T Ω R 3 The system Results T d 2 p f p p + P( e θ v 2 f ), and if p =, sup d 2 p f (t) T d 2 p f + P( e θ v 2 f ), where d 2 p = d 2 p (t, x, v) (defined later) vanishes on the grazing boundary γ.

New results: regularity in [GKTT 13] Assume Ω is a smooth strictly convex domain. Assume (H ) and f satisfies the diffuse boundary condition. Theorem (Weighted, p [2, ]) Assume d 2 p f L p (Ω R 3 ). Then T < T, if p [2, [ T sup dp 2 f p p(t) + dp 2 f p v.n ds dv dt t T Ω R 3 The system Results T d 2 p f p p + P( e θ v 2 f ), and if p =, sup d 2 p f (t) T d 2 p f + P( e θ v 2 f ), where d 2 p = d 2 p (t, x, v) (defined later) vanishes on the grazing boundary γ. Theorem (C 1 propagation) If d 2 f C ( Ω R 3 ) and tf satisfies the diffuse boundary condition, then f is C 1 away from the grazing boundary.

New results: regularity in non- [GKTT 14] Assume Ω is a smooth bounded domain. Assume (H ). Theorem (BV propagation) If f BV (Ω R 3 ), then T < T, The system Results sup f (t) BV T f BV + P( e θ v 2 f ), t T and x,v f γ is a Radon measure σ(t) on Ω R 3 such that T σ( Ω R 3 ) dt T f BV + P( e θ v 2 f ).

The system Results Convex domains Non-

Road map for a priori estimates 1. for f (uses convexity) T sup f (t) p p + f p γ +,p f p p + I Ω (T ) + J Ω (T ). t T 2. Estimates inside the domain { } J Ω T sup s t f p p + P( e θ v 2 f ). 3. Estimates on the boundary (uses p > 1) T I Ω O(ε) f p dγds + tp( e θ v 2 f ) γ + [ T ] +C ε,t f p p + J Ω + f (s) p p. We conclude choosing ε 1, T 1.

Computation of the traces: a model problem Consider the system th + v xh + νh = H, h t= = f, h γ = g. Define tf and xg = ( τ1 g, τ2 g, ng) by tf := v xf νf + H t=, ng := 1 { tg } (v τ i ) τi g νg + H. v n Proposition If H C ([, T ], L p (Ω R 3 )), f L p (Ω R 3 ) and g C ([, T ], L p (γ )), if H, f, g strongly decay in v and g() = f γ, then h t= = f and h γ = g.

Computation of the traces: characteristics Proof: solve the system explicitly (method of characteristics). t> t-t b Ω x b (x,v) If t > t b, h(t, x, v) = e t b ν g(t t b, x b, v) + tb e s ν H(t s, x sv, v) ds

Computation of the traces: characteristics Remark: singularity in non-. Ω (x,v) (x',v) x' b x b If t > t b, h(t, x, v) = e t b ν g(t t b, x b, v) + tb e s ν H(t s, x sv, v) ds

For [ t, x, v ] t f + v x f + ν[ µf ] f = Γ(f, f ) + v xf + f ν[ µf ], f t= = f, { µ(v) } f γ = µ(u)f (u)u n(x) du. u n(x)>

For [ t, x, v ] t f + v x f + ν[ µf ] f = Γ(f, f ) + v xf + f ν[ µf ], f t= = f, { µ(v) } f γ = µ(u)f (u)u n(x) du. u n(x)> : f (t) p p + + t t f p γ +,p + t ν 1/p f p p = f p p + t f p γ,p Ω R 3 { Γ(f, f ) + v xf + f ν[ µf ]} f p 2 f. where dγ = v n(x) dx dv and γ+,p = L p (γ +, dγ).

Road map for a priori estimates 1. for f (uses convexity) T sup f (t) p p + f p γ +,p f p p + I Ω (T ) + J Ω (T ). t T 2. Estimates inside the domain { } J Ω T sup s t f p p + P( e θ v 2 f ). 3. Estimates on the boundary (uses p > 1) T I Ω O(ε) f p dγds + tp( e θ v 2 f ) γ + [ T ] +C ε,t f p p + J Ω + f (s) p p. We conclude choosing ε 1, T 1.

We use Grad s estimates [Grad 62] to show for any h = h(t, x, v) Therefore J Ω = Γ(h, h) p e θ v 2 h [ h p + h p], h ν[ µh] p e θ v 2 h [ h p + h p]. T T { Γ(f, f ) + v xf + f ν[ µf ]} f p 2 f Ω R 3 Γ(f, f ) + v xf + f ν[ µf ] p f p 1 p P( sup e θ v 2 f ) t T { T } T + f p p.

Road map for a priori estimates 1. for f (uses convexity) T sup f (t) p p + f p γ +,p f p p + I Ω (T ) + J Ω (T ). t T 2. Estimates inside the domain J Ω P( sup e θ v 2 f ) t T { T } T + f p p. 3. Estimates on the boundary (uses p > 1) T I Ω O(ε) f p dγds + tp( e θ v 2 f ) γ + [ T ] +C ε,t f p p + J Ω + f (s) p p. We conclude choosing ε 1, T 1.

Diffuse boundary f γ = c µ µ(v) tf γ = c µ µ(v) τ f γ = c µ µ(v) + c µ µ(v) u n(x)> u n(x)> u n(x)> u n(x)> µ(u) f (u)u n(x) du. µ(u) tf (u)u n(x) du, µ(u) τ f (u)u n(x) du, µ(u) τ T t (x)t (x)u f (u)u n(x) du, nf γ = 1 { tf } (v τ i ) τi f νf + Γ(f, f ), v n v f γ = c µv µ(v) µ(u) f (u)u n(x) du. u n(x)>

Diffuse boundary f γ = c µ µ(v) tf γ = c µ µ(v) τ f γ = c µ µ(v) + c µ µ(v) u n(x)> u n(x)> u n(x)> u n(x)> µ(u) f (u)u n(x) du. µ(u) tf (u)u n(x) du, µ(u) τ f (u)u n(x) du, µ(u) τ T t (x)t (x)u f (u)u n(x) du, nf γ = 1 { tf } (v τ i ) τi f νf + Γ(f, f ), v n v f γ = c µv µ(v) µ(u) f (u)u n(x) du. u n(x)> f γ ( 1 + v n 1) µ(v) 1/4 u n(x)> µ(u) f (u) u n du + O(f )

Diffuse boundary f γ p O(f ) + ( 1 + v n(x) p) µ(v) p/4 p µ(u) f (u) u n(x) du. We integrate T u n(x)> f p v n(x) dv dx dt TO(f ) γ ( + 1 + v n 1 p ) µ(v) p/8 dv R 3 T p µ(u) f (u) u n(x) du dx dt. Ω u n(x)> To control the singularity: p < 2.

Diffuse boundary f γ p O(f ) + ( 1 + v n(x) p) µ(v) p/4 p µ(u) f (u) u n(x) du. We integrate T u n(x)> f p v n(x) dv dx dt TO(f ) γ ( + 1 + v n 1 p ) µ(v) p/8 dv R 3 T p µ(u) f (u) u n(x) du dx dt. Ω u n(x)> To control the singularity: p < 2.

Outgoing flow Idea: the outgoing flow is determined by "inside" quantities during the time [t t b, t]. t> t-s, <s<t b t-t b x b=x-vt b (x-vs,v) (x,v) Ω This is not satisfying when t b 1, i.e. on the almost grazing boundary γ ɛ + = {(x, v) Ω R 3 : < v n(x) < ɛ or v 1/ɛ}.

Outgoing flow Idea: the outgoing flow is determined by "inside" quantities during the time [t t b, t]. t> t-s, <s<t b t-t b x b=x-vt b (x-vs,v) (x,v') Ω (x,v) x' b This is not satisfying when t b 1, i.e. on the almost grazing boundary γ ɛ + = {(x, v) Ω R 3 : < v n(x) < ɛ or v 1/ɛ}.

Outgoing flow: grazing and non-grazing boundary Away from the almost grazing boundary, p µ(u) 1/2 f (u) u n(x) du dx dt Ω u:(x,u) γ +\γ+ ε [ T ] C ε,t f p p + J Ω + f (s) p p.

Outgoing flow: grazing and non-grazing boundary Away from the almost grazing boundary, p µ(u) 1/2 f (u) u n(x) du dx dt Ω u:(x,u) γ +\γ+ ε [ T ] C ε,t f p p + J Ω + f (s) p p. For the almost grazing part, by Hölder s inequality p µ(u) 1/2 f (u) u n(x) du dx dt Ω u:(x,u) γ+ ε O(ε) f p dγ. γ +

Outgoing flow: grazing and non-grazing boundary Away from the almost grazing boundary, p µ(u) 1/2 f (u) u n(x) du dx dt Ω u:(x,u) γ +\γ+ ε [ T ] C ε,t f p p + J Ω + f (s) p p. For the almost grazing part, by Hölder s inequality p µ(u) 1/2 f (u) u n(x) du dx dt Ω u:(x,u) γ+ ε O(ε) f p dγ. γ + Remark: this computation breaks at p = 1!

Road map for a priori estimates 1. for f (uses convexity) T sup f (t) p p + f p γ +,p f p p + I Ω (T ) + J Ω (T ). t T 2. Estimates inside the domain J Ω P( sup e θ v 2 f ) t T { T } T + f p p. 3. Estimates on the boundary (uses p > 1) T I Ω O(ε) f p dγds + TP( sup e θ v 2 f ) γ + t T [ T ] +C ε,t f p p + J Ω + f (s) p p.

Road map for a priori estimates 1. for f (uses convexity) T sup f (t) p p + f p γ +,p f p p + I Ω (T ) + J Ω (T ). t T 2. Estimates inside the domain J Ω P( sup e θ v 2 f ) t T { T } T + f p p. 3. Estimates on the boundary (uses p > 1) T I Ω O(ε) f p dγds + TP( sup e θ v 2 f ) γ + t T [ T ] +C ε,t f p p + J Ω + f (s) p p. To conclude: recall sup t T e θ v 2 f e θ v 2 f, choose ε 1 and use a Gronwall s argument.

The system Results Convex domains Non- Convex domains Non-

Weighted in The (strictly convex) domain Ω is defined by Ω = {x R 3 : ξ(x) < }. Define the kinetic distance α(x, v) := v xξ 2 2{v 2 xξv}ξ(x). Convex domains Non- Main features: "distance": vanishes exactly on the grazing boundary "kinetic": invariant along the characteristics (up to some quantity in v ).

Weighted in The (strictly convex) domain Ω is defined by Ω = {x R 3 : ξ(x) < }. Define the kinetic distance α(x, v) := v xξ 2 2{v 2 xξv}ξ(x). Convex domains Non- Main features: "distance": vanishes exactly on the grazing boundary "kinetic": invariant along the characteristics (up to some quantity in v ). We therefore consider, for ω 1, d 2 p (t, x, v) := e ω<v>t α(x, v) βp.

W 1, propagation in No energy identity: follow the stochastic backward characteristics and control the number of "bounces"... Ω Convex domains Non- (x,v ) 1 1 (x,v) (x,v ) 3 3 (x,v ) 4 4 (x,v ) 2 2

BV propagation in non- We get rid of the grazing trajectories S B := {(x, v) Ω R 3 : n(x b (x, v)) v = } by applying a cut-off on a tubular neigborhood of S B (much work!). Convex domains Non- Ω For the W 1,1 propagation: to estimate the grazing boundary term we use the boundary condition twice (double iteration scheme).

BV propagation in non- We get rid of the grazing trajectories S B := {(x, v) Ω R 3 : n(x b (x, v)) v = } by applying a cut-off on a tubular neigborhood of S B (much work!). Convex domains Non- Ω For the W 1,1 propagation: to estimate the grazing boundary term we use the boundary condition twice (double iteration scheme).

BV propagation in non- We get rid of the grazing trajectories S B := {(x, v) Ω R 3 : n(x b (x, v)) v = } by applying a cut-off on a tubular neigborhood of S B (much work!). x b (x,v) Ω Convex domains Non- For the W 1,1 propagation: to estimate the grazing boundary term we use the boundary condition twice (double iteration scheme).

The system Results Convex domains Non-

Recap: propagation of regularity In In non- C away from γ [Guo 1] Discontinuity created on γ and propagated along the W 1,p for 1 < p < 2 grazing trajectories [Kim 11] Ct-ex. (transport eq.) : xf / L 2 BV propagation Weighted W 1,p for p [2, ] C 1 away from γ Ct-ex. : 2 f / L 1

Further results Other boundary conditions [GKTT13] In, propagation of C 1 regularity away from γ (with the help of the kinetic distance) for the specular and bounce-back boundary conditions. Non-isothermal boundary [EGKM13] Results of existence, uniqueness and stability (exponential convergence towards the solution of the stationnary problem) for a (non too much) varying boundary temperature. Continuity propagation in convex domains.

Esposito, R.; Guo, Y.; Kim, C. ; Marra, R: Non-Isothermal Boundary in the Boltzmann Theory and Fourier Law, Comm. Math. Phys. 323 (213) 177 239. Grad, H.: Asymptotic theory of the Boltzmann equation, II. Rarefied gas dynamics, Proceedings of the 3rd international Symposium 26 59, Paris, (1962). Guo, Y.: Decay and Continuity of in Bounded Domains. Arch. Rational Mech. Anal. 197 (21) 713 89. Guo, Y.; Kim, C.; Tonon, D.; T. A.: Boltzmann Equation in Convex Domains. Submitted. Guo, Y.; Kim, C.; Tonon, D.; T. A.: BV-regularity of the in Non-Convex Domains. Submitted. Kim, C.: Formation and propagation of discontinuity for Boltzmann equation in non-. Comm. Math. Phys. 38 (211) 641 71.