Loading [MathJax]/jax/output/SVG/jax.js
Research article Special Issues

On nonlinear Schrödinger equations with random initial data

  • This note is concerned with the global well-posedness of nonlinear Schrödinger equations in the continuum with spatially homogeneous random initial data.

    Citation: Mitia Duerinckx. On nonlinear Schrödinger equations with random initial data[J]. Mathematics in Engineering, 2022, 4(4): 1-14. doi: 10.3934/mine.2022030

    Related Papers:

    [1] José Antonio Vélez-Pérez, Panayotis Panayotaros . Wannier functions and discrete NLS equations for nematicons. Mathematics in Engineering, 2019, 1(2): 309-326. doi: 10.3934/mine.2019.2.309
    [2] Luis Silvestre, Stanley Snelson . Solutions to the non-cutoff Boltzmann equation uniformly near a Maxwellian. Mathematics in Engineering, 2023, 5(2): 1-36. doi: 10.3934/mine.2023034
    [3] Michele Dolce, Ricardo Grande . On the convergence rates of discrete solutions to the Wave Kinetic Equation. Mathematics in Engineering, 2024, 6(4): 536-558. doi: 10.3934/mine.2024022
    [4] Helen Christodoulidi, Christos Efthymiopoulos . Stages of dynamics in the Fermi-Pasta-Ulam system as probed by the first Toda integral. Mathematics in Engineering, 2019, 1(2): 359-377. doi: 10.3934/mine.2019.2.359
    [5] Gabriel B. Apolinário, Laurent Chevillard . Space-time statistics of a linear dynamical energy cascade model. Mathematics in Engineering, 2023, 5(2): 1-23. doi: 10.3934/mine.2023025
    [6] Giuseppe Maria Coclite, Lorenzo di Ruvo . On the initial-boundary value problem for a Kuramoto-Sinelshchikov type equation. Mathematics in Engineering, 2021, 3(4): 1-43. doi: 10.3934/mine.2021036
    [7] Iakovidis Isidoros, Nicola Arcozzi . Improved convergence rates for some kernel random forest algorithms. Mathematics in Engineering, 2024, 6(2): 305-338. doi: 10.3934/mine.2024013
    [8] Diogo Gomes, Julian Gutierrez, Ricardo Ribeiro . A mean field game price model with noise. Mathematics in Engineering, 2021, 3(4): 1-14. doi: 10.3934/mine.2021028
    [9] E. Artioli, G. Elefante, A. Sommariva, M. Vianello . Homogenization of composite materials reinforced with unidirectional fibres with complex curvilinear cross section: a virtual element approach. Mathematics in Engineering, 2024, 6(4): 510-535. doi: 10.3934/mine.2024021
    [10] Carlo Ciliberto, Massimiliano Pontil, Dimitrios Stamos . Reexamining low rank matrix factorization for trace norm regularization. Mathematics in Engineering, 2023, 5(3): 1-22. doi: 10.3934/mine.2023053
  • This note is concerned with the global well-posedness of nonlinear Schrödinger equations in the continuum with spatially homogeneous random initial data.



    Motivated by weak turbulence theory, e.g., [10], we consider nonlinear Schrödinger equations with spatially homogeneous statistical ensembles of initial data. As a prototypical example, we study the defocusing cubic equation in Rd,

    itu=u+|u|2u,u|t=0=u. (1.1)

    A statistical ensemble of initial data amounts to considering an initial condition that is a realization u(,ω) of a random field u:Rd×ΩC on some probability space (Ω,P). The spatial homogeneity condition then implies that initial mass and energy diverge:*

    *Under an additional ergodicity assumption, it further entails Rd|u(,ω)|2=Rd|u(,ω)|2+12|u(,ω)|4= almost surely.

    E[Rd|u|2]=,E[Rd|u|2+12|u|4]=.

    This divergence is a key aspect at the very core of weak turbulence: Strichartz' estimates are not applicable in this infinite-energy setting, which is thus in sharp contrast with the finite-energy phenomenology and scattering results [7].

    The present note is concerned with the global well-posedness of (1.1) in this infinite-energy setting. The main difficulty is related to the lack of a uniform bound on the propagation speed: mass that is initially spread out might move together and blow up. This contrasts with the case of the nonlinear wave equation, as well as of the discrete nonlinear Schrödinger equation, for which there is an (approximate) finite propagation speed and global well-posedness follows, see [3,Propositions 1-3]. As explained in Examples 2.3 below, periodic and quasi-periodic initial data can in fact be viewed as particular instances of the spatially homogeneous random setting. While the periodic case is well understood [2], the almost periodic case remains largely open and we refer to recent work by Oh [11,12] on the topic. In the general random setting, the problem seems to have only been considered very recently by Dodson, Soffer, and Spencer [3], who established local well-posedness in the real analytic category. If the nonlinearity |u|2u in (1.1) is replaced by a regularized version |ϕu|2(ϕu) for some smooth decaying kernel ϕ, then the problem is strongly reduced and global well-posedness is obtained in [3] in the Ck category.

    Our main result in this note states the global well-posedness of the nonlinear Schrödinger equation (1.1) in the spatially homogeneous energy space provided that a tiny dissipation is added. This tiny dissipation is physically relevant in the context of weak turbulence, e.g., [10], and the constructed solution is controlled uniformly with respect to this dissipation. The definition of a meaningful vanishing-dissipation limit remains an open problem (beyond the local-in-time real analytic framework of [3]). Precise definitions of spatial homogeneity and of the functional space X below are postponed to the next section.

    Theorem 1. Let 1d<4. Given a probability space (Ω,P), let X be the Banach space of spatially homogeneous jointly measurable random fields v:Rd×ΩC with

    vX:=E[|v|2]12+E[|v|4]14<.

    For all ε>0 and uX, there exists a unique global weak solution uεL(R+;X) to the equation

    (ε+i)tuε=uε+|uε|2uε,uε|t=0=u, (1.2)

    in the sense that Duhamel's formula holds almost everywhere,

    utε=etεiu1εit0etsεi(|usε|2usε)ds,t0.

    In addition, it satisfies the following dissipation estimates: for all t0,

    E[|utε|2]E[|utε|2]+2ε1+ε2t0E[|usε|2+|usε|4]ds=E[|u|2],E[|utε|2+12|utε|4]E[|u|2+12|u|4].

    Notation.

    ● We denote by C1 any constant that only depends on the space dimension d. We use the notation (resp. ) for C× (resp. 1C×) up to such a multiplicative constant C. We write when both and hold. We add subscripts to C,,, to indicate dependence on other parameters.

    ● The ball centered at x and of radius r in Rd is denoted by Br(x), and we write for abbreviation B(x):=B1(x) and Br:=Br(0).

    Given a reference probability space (Ω,P), we recall the notion of statistical spatial homogeneity for random fields.

    Definition 2.1. A random field on Rd is a map v:Rd×ΩC such that for all xRd the function v(x,):ΩC is measurable. It is said to be (statistically) spatially homogeneous if its finite-dimensional law is shift-invariant, that is, if for any finite set ERd the law of {v(x+y,)}xE does not depend on the shift yRd. In addition, it is said to be jointly measurable if the map v:Rd×ΩC is jointly measurable. We denote by L0hom(Rd×Ω) the set of spatially homogeneous jointly measurable random fields.

    Note that the joint measurability condition ensures that realizations v(,ω) are almost surely measurable functions on Rd and can thus be taken as meaningful initial data in (1.1) or (1.2). The following result by von Neumann [14] gives an alternative characterization of joint measurability in this context, which can be viewed as a stochastic version of Lusin's theorem (see also [8,Section 7.1]).

    Lemma 2.2 (Joint measurability; [8,14]). A spatially homogeneous random field v is jointly measurable if and only if it is stochastically continuous, that is, if for all δ>0 it satisfies

    P[|v(x,)v(y,)|>δ]0as|xy|0.

    Examples 2.3. Important examples of spatially homogeneous random fields are found among Gaussian fields, and we also explain how periodic and almost periodic settings can be viewed as particular instances of this random framework (see also [13,p.846]).

    (a) Gaussian fields: A gauge-invariant Gaussian random field v is a family {v(x,)}xRd of complex-valued Gaussian random variables such that v and eiθv have the same finite-dimensional law for all θ[0,2π). Equivalently, this means for all x,y,

    E[v(x,)]=0,Cov[v(x,);v(y,)]=0,

    and we denote by c(x,y):=Cov[¯v(x,);v(y,)] the covariance function. This random field v is spatially homogeneous if and only if c is of the form c(x,y)=c0(xy) for some function c0:RdC. Note that c0 is necessarily a positive definite bounded function. In addition, the field v is stochastically continuous, hence jointly measurable by Lemma 2.2, if and only if c0 is continuous at the origin.

    (b) Periodic setting: Given a 1-periodic measurable function vper:RdC, we choose the probability space (Ω,P) as the periodic cell [0,1)d endowed with Lebesgue's measure, and we define an associated random field v:Rd×ΩC by v(x,ω):=vper(x+ω). The latter is clearly spatially homogeneous and jointly measurable, and for ω=0 we recover v(x,0)=vper(x).

    (c) Almost periodic setting: Denote by B(Rd) the Bohr compactification of the additive group (Rd,+) and let b:RdB(Rd) the associated continuous homomorphism, see e.g., [9]. By definition, given an almost periodic function vap:RdC, there exists a continuous function Vap:B(Rd)C such that vap=Vapb. We choose the probability space (Ω,P) as B(Rd) endowed with its normalized Haar measure, and we define a random field v:Rd×ΩC by v(x,ω):=Vap(b(x)+ω), where we use the notation '+' for the group law on B(Rd). This random field is clearly spatially homogeneous and jointly measurable, and for ω=0 we recover v(x,0)=vap(x).

    In this section, we define more carefully the functional space X used in the statement of Theorem 1. For 1q<, we denote by Lqhom(Rd×Ω) the Banach space of spatially homogeneous jointly measurable random fields vL0hom(Rd×Ω) such that the following norm is finite,

    vLqhom(Rd×Ω):=vLq([0,1)d×Ω).

    By spatial homogeneity, see Definition 2.1, this is in fact equivalent to

    vLqhom(Rd×Ω)=v(x,)Lq(Ω)for any xRd.

    As Lqhom(Rd×Ω) is invariant under spatial translations and as its elements are almost surely locally Lq-integrable, the spatial gradient can be defined on Lqhom(Rd×Ω) and its domain is denoted by W1,qhom(Rd×Ω). More generally, for all s0, we define Ws,qhom(Rd×Ω) as the Banach space of random fields vLqhom(Rd×Ω) such that the following norm is finite,

    vWs,qhom(Rd×Ω):=(1)s2vLq([0,1)d×Ω),

    and for q=2 we use the usual notation Hshom(Rd×Ω):=Ws,2hom(Rd×Ω). In these terms, the space X in Theorem 1 coincides with H1homL4hom(Rd×Ω).

    Next, we give an alternative description of these spaces and we explain how equations (1.1) or (1.2) in this spatially homogeneous random setting are equivalent to abstract equations on the probability space; this construction is standard for corrector equations in stochastic homogenization theory, see e.g., [13,Section 2] and [8,Section 7.1]. Let uL0hom(Rd×Ω) be a reference random field. Since we consider equations with realizations of u as initial data, we can henceforth assume that the probability space (Ω,P) is endowed with the σ-algebra σ(u) generated by u. Translations u(,ω)u(+x,ω) then induce a unique multiplicative linear action T={Tx}xRd of the additive group (Rd,+) on the algebra of random variables.

    That is, the σ-algebra generated by all sets of the form {ωΩ:u(x1,ω)A1,,u(xn,ω)An} with n1, x1,,xnRd, and Borel subsets A1,AnC.

    Lemma 2.4 (Properties of T).

    (i) For 1q, the maps Tx's are isometries on Lq(Ω).

    (ii) For 1q<, the action T is a C0-group of isometries on Lq(Ω).

    Proof. Item (i) follows from the fact that u is spatially homogeneous. We turn to the proof of (ii). Given q<, it remains to check that TxXXLq(Ω)0 as |x|0 for all XLq(Ω). By a truncation argument, it suffices to argue for XL(Ω). By the joint measurability of u, the map (x,ω)(TxX)(ω) is also jointly measurable, hence stochastically continuous by Lemma 2.2. Writing for any δ>0,

    TxXXqLq(Ω)δq+(2XL(Ω))qP[|TxXX|>δ],

    the conclusion follows from stochastic continuity.

    In terms of T, we can define the extension XLqhom(Rd×Ω) of any random variable XLq(Ω), and the restriction vLq(Ω) of any random field vLqhom(Rd×Ω), via

    X(x,ω):=(TxX)(ω),v(ω)=v(0,ω),

    and we note that (X)=X and (v)=v, thus yielding a canonical isomorphism

    Lq(Ω)Lqhom(Rd×Ω). (2.1)

    For 1q<, as T is a C0-group of isometries on Lq(Ω), cf. Lemma 2.4(ii), we can define the T-gradient as the generator of this group. It is a densely defined operator on Lq(Ω), its domain is denoted by W1,q(Ω), and it is skew-adjoint on L2(Ω). Alternatively, this operator can be reinterpreted via the isomorphism (2.1):

    W1,q(Ω)W1,qhom(Rd×Ω),(X)=Xfor all XW1,q(Ω).

    We also define the corresponding T-Laplacian := on Lq(Ω), which is nonnegative and essentially self-adjoint on L2(Ω). For all s0, we denote by Ws,q(Ω) the Banach space of random variables XLq(Ω) such that XWs,qhom(Rd×Ω).

    The above construction entails that spatially homogeneous solutions of (1.2) are equivalent to solutions of a corresponding abstract equation on the probability space. Note that expressions like etεiv in (2.2) below make sense almost surely for vLqhom(Rd×Ω) since the kernel of etεi has Gaussian decay while realizations of v have subexponential growth almost surely, cf. (3.13). In the periodic case, in view of Example 1(b), is the periodic gradient and this result amounts to reducing (1.2) to the corresponding equation on the periodic cell Ω=[0,1)d.

    Lemma 2.5. Given T>0, the following two properties are equivalent:

    A random field uεL([0,T];H1homL4hom(Rd×Ω)) is a weak solution of (1.2) in the sense that Duhamel's formula holds almost everywhere,

    utε=etεiu1εit0etsεi(|usε|2usε)ds,0tT. (2.2)

    We have utε=(Utε) where UεL([0,T];H1L4(Ω)) is a weak solution of the following abstract equation on the probability space,

    (ε+i)tUε=Uε+|Uε|2Uε,Uε|t=0=(u), (2.3)

    in the sense that Duhamel's formula holds almost everywhere,

    Utε=etεi(u)1εit0etsεi(|Usε|2Usε)ds,0tT.

    In contrast with the periodic case, the T-Laplacian on L2(Ω) typically has absolutely continuous spectrum and no spectral gap above 0. As shown in [4], the spectrum can actually be arbitrary and depends on the structure of the underlying probability space; we focus here for simplicity on the Gaussian setting.

    Lemma 2.6 (Spectrum of T-Laplacian; [4]). Assume that uL2hom(Rd×Ω) is Gaussian in the sense of Example 1(a) and that its covariance function c0 has an absolutely continuous Fourier transform. Then, the spectrum of on L2(Ω) is [0,) and is made of a simple eigenvalue at 0 embedded in absolutely continuous spectrum.

    In particular, this entails that Poincaré's inequality and compact Rellich embeddings do not hold on H1(Ω). In addition, we show that Sobolev embeddings also fail and that the parabolic semigroup {et}t0 yields no improved integrability. Heuristically, this lack of functional tools is related to the fact that the T-gradient only contains information on a finite set of directions, while Ω is typically an infinite product space. This constitutes a key difficulty for the analysis of nonlinear equations such as (1.2) in this setting.

    Lemma 2.7 (Lack of functional tools). Assume that uL2hom(Rd×Ω) is Gaussian in the sense of Example 1(a) and that its covariance function c0 is integrable. Then,

    (i) Poincaré's inequality: XE[X]L2(Ω)CXL2(Ω) does not hold on H1(Ω).

    (ii) Compact Rellich embedding: Hm(Ω) is not compactly embedded in L2(Ω) for any m>0.

    (iii) Sobolev embedding: XLq(Ω)CXHm(Ω) does not hold on Hm(Ω) for any m0 and q>2.

    (iv) Parabolic improvement of integrability: ezXLq(Ω)CXL2(Ω) does not hold on Lq(Ω) for any q>2 and any zC with z0.

    Proof. We start with items (i) and (ii). They can both be viewed as consequences of Lemma 2.6, but we rather provide a quick direct proof. Given a real-valued test function ξCc(Rd), consider the Gaussian random variables

    Xn(ω):=nd2Rdξ(yn)u(y,ω)dy,n1.

    A direct computation yields as n,

    Xn2L2(Ω)R:=ξ2L2(Rd)Rdc0,n2Xn2L2(Ω)ξ2L2(Rd)Rdc0, (2.4)

    which contradicts the validity of Poincaré's inequality, hence proves (i). Since the sequence (Xn)n is bounded in Hm(Ω) for any m0, and since it converges weakly but not strongly to 0 in L2(Ω), item (ii) follows.

    We turn to the proof of item (iii). A simple application of Wick's theorem yields for all integers r1,

    Xn2rL2r(Ω)=r!Xn2rL2(Ω).

    Extending this to all r1, and combining with (2.4), we find

    Xn2rL2r(Ω)nRrΓ(r+1). (2.5)

    Also note that (2.4) yields mXn2rL2r(Ω)0 for all r,m1. Hence, given q2 and m0, we deduce for all integers p1,

    XpnLq(Ω)XpnHm(Ω)nΓ(qp+1)1/qΓ(2p+1)1/2.

    For q>2, the right-hand side blows up exponentially as p, thus contradicting the validity of Sobolev's inequality.

    It remains to prove item (iv) and we start with the easy case when z=0. The essential self-adjointness of the T-Laplacian on L2(Ω) entails that ez is unitary on L2(Ω) in that case, hence the inequality ezXLq(Ω)CXL2(Ω) for all XL2(Ω) would in fact imply XLq(Ω)CXL2(Ω) for all XL2(Ω), a contradiction.

    We turn to the proof of (iv) for z>0. As the operator ez has kernel Kz(x):=(4πz)d2e14z|x|2, and as realizations of the Gaussian random field Xn have subexponential growth almost surely, we can write in view of (2.1),

    ezXpn=RdKz(x)Xn(x,)pdx.

    Appealing to Wick's theorem as above, using (2.4) in form of

    E[Xn(x,)¯Xn(y,)]nR,

    and noting that RdKz=1, we find for all r1 and all integers p1,

    ezXpn2rL2r(Ω)nRprΓ(pr+1).

    Hence, combined with (2.5), given q2, we deduce for all integers p1,

    ezXpnLq(Ω)XpnL2(Ω)nΓ(12qp+1)1/qΓ(p+1)1/2.

    For q>2, the right-hand side blows up exponentially as p, and the conclusion follows.

    We turn to the proof of Theorem 1, where we recall X=H1homL4hom(Rd×Ω) with the above notation. In order to overcome the lack of functional tools to study the nonlinear equation (1.2) in this setting, cf. Lemma 2.7, we rather focus on almost sure realizations in local Sobolev spaces, for which standard tools are available. Dissipation is crucial to compensate for the lack of finite propagation speed and allows to prove well-posedness in local Sobolev spaces with polynomial growth, which is then post-processed into a well-posedness result in the spatially homogeneous random setting.

    Proof of Theorem 1. We start by introducing local Sobolev spaces with polynomial growth, which are natural spaces to control almost sure realizations of random fields. Given 0, we consider the uniformly localized Lq and H1 norms with -growth,

    vLquloc,(Rd):=supx0Rd(x0qB(x0)|v|q)1q,vH1uloc,(Rd):=supx0Rd(x02B(x0)(|v|2+|v|2))12,

    and we denote by Lquloc,(Rd) and H1uloc,(Rd) the corresponding subspaces of L1loc(Rd). We also consider the uniformly localized energy functional with -growth,

    Euloc,(v):=supx0Rd(x02B(x0)(|v|2+12|v|4)).

    As shown below, cf. (3.13), given >dq, realizations of a random field vLqhom(Rd×Ω) belong almost surely to Lquloc,(Rd). With this in mind, we start by studying the nonlinear Schrödinger equation (1.2) in H1uloc,(Rd), and next we exploit uniqueness to construct a unique dynamics in H1homL4hom(Rd×Ω). Note that H1uloc,(Rd) embeds in L4uloc,(Rd) by the Sobolev embedding (with d4), while on the contrary H1hom(Rd×Ω) does in general not embed in L4hom(Rd×Ω), cf. Lemma 2.7(iii). The proof is split into two main steps.

    Step 1. Global well-posedness in H1uloc,(Rd).

    For d<4, given ε>0 and 0, we show that for all vH1uloc,(Rd) the equation

    (ε+i)tvε=vε+|vε|2vε,vε|t=0=v

    admits a unique global weak (Duhamel) solution vε in Lloc(R+;H1uloc,(Rd)). We split the proof into four further substeps.

    Substep 1.1. Equivalent definition of Lquloc,(Rd).

    In terms of the exponential cut-off χ(x):=e|x|, setting for abbreviation χx0:=χ(x0), we show that for all q, we have

    vLquloc,(Rd)q,supx0Rd(x0qRdχx0|v|q)1q. (3.1)

    Indeed, as χx01 on B(x0), the left-hand side in (3.1) is clearly bounded above by the right-hand side. The converse inequality follows from the following,

    (x0qRdχx0|v|q)1q(x0qzZdB(x0+z)χx0|v|q)1q(x0qzZde|z|x0+zq)1qvLquloc,(Rd)vLquloc,(Rd).

    Substep 1.2. Localized parabolic estimates.

    Given ε>0, 0, and 1pq with d2(1p1q)<1, we show for all gLquloc,(Rd) and t0,

    etεigLquloc,(Rd)p,q,,εtd2(1p1q)eCεtgLpuloc,(Rd), (3.2)

    and in addition, provided d2(1p1q)<12,

    etεigLquloc,(Rd)p,q,,εtd2(1p1q)12eCεtgLpuloc,(Rd). (3.3)

    Multiplying the parabolic evolution {etεig}t0 with the exponential cut-off χx0, and using Duhamel's formula, we easily find

    χx0(etεig)=etεi(χx0g)+1εit0etsεi((χx0)(esεig))ds2εit0etsεi((χx0)(esεig))ds. (3.4)

    Note that the parabolic semigroup {etεi}t0 has kernel Ktε(x):=(εi4πt)d2eεi4t|x|2, which implies by Young's convolution inequality, for all gCc(Rd) and 1pq, letting r be such that 1p+1r=1q+1,

    etεigLq(Rd)=KtεgLq(Rd)KtεLr(Rd)gLp(Rd)p,q,εtd2(1p1q)gLp(Rd), (3.5)

    and similarly,

    etεigLq(Rd)p,q,εtd2(1p1q)12gLp(Rd). (3.6)

    For 1pq, taking the Lq norm in (3.4) and using these parabolic estimates, we obtain

    χx0(etεig)Lq(Rd)Cp,q,εtd2(1p1q)χx0gLp(Rd)+Cεt0(1+(ts)12)χx0(esεig)Lq(Rd)ds,

    and the claim (3.2) easily follows from Grönwall's inequality together with (3.1). Using (3.6) instead of (3.5), the claim (3.3) is obtained in a similar way.

    Substep 1.3. Local well-posedness in Lquloc,(Rd).

    Given ε>0, 0, and 3q with q>d, we show that for all vLquloc,(Rd) there exists T>0 such that the equation

    (ε+i)tvε=vε+|vε|2vε,vε|t=0=v

    admits a unique weak (Duhamel) solution vε in L([0,T];Lquloc,(Rd)). In case q>6dd+2, we further have vεL([0,T];H1uloc,(Rd)) provided vH1uloc,(Rd).

    To prove well-posedness in L([0,T];Lquloc,(Rd)), we argue by a Picard fixed-point argument: for T>0 we define an operator ΦT,ε(;v) on L([0,T];Lquloc,(Rd)) by

    (ΦT,ε(v;v))t:=etεiv1εit0etsεi(|vs|2vs)ds,0tT, (3.7)

    and it suffices to show that for all v,wL([0,T];Lquloc,(Rd)),

    ΦT,ε(v;v)L([0,T];Lquloc,(Rd))q,,εeCεT(vLquloc,(Rd)+v3L([0,T];Lquloc,(Rd))), (3.8)
    ΦT,ε(v;v)ΦT,ε(w;v)L([0,T];Lquloc,(Rd))q,,εT1dqeCεT(v,w)2L([0,T];Lquloc,(Rd))vwL([0,T];Lquloc,(Rd)). (3.9)

    It remains to prove these two estimates. Applying the localized parabolic estimate (3.2) with exponents q3q, we find for all 3q with q>d,

    (ΦT,ε(v;v))tLquloc,(Rd)q,,εeCεt(vLquloc,(Rd)+t0(ts)dq|vs|2vsLq/3uloc,(Rd)ds)q,,εeCεt(vLquloc,(Rd)+t1dqv3L([0,t];Lquloc,(Rd))),

    which proves (3.8). Similarly, we find

    (ΦT,ε(v;v))t(ΦT,ε(w;v))tLquloc,(Rd)q,,εeCεtt0(ts)dq|vs|2vs|ws|2wsLq/3uloc,(Rd)ds,

    and thus, further using Hölder's inequality in form of

    |vs|2vs|ws|2wsLq/3uloc,(Rd)(vs,ws)2Lquloc,(Rd)vswsLquloc,(Rd),

    we deduce

    (ΦT,ε(v;v))t(ΦT,ε(w;v))tLquloc,(Rd)q,,εt1dqeCεt(v,w)2L([0,t];Lquloc,(Rd))vwL([0,t];Lquloc,(Rd)),

    which proves (3.9).

    It remains to show that for q>6dd+2 this local weak (Duhamel) solution vε=ΦT,ε(vε;v) in L([0,T];Lquloc,(Rd)) also belongs to L([0,T];H1uloc,(Rd)) provided vH1uloc,(Rd). Without loss of generality we choose q6. As the condition on q ensures d2(3q12)<12, we appeal to the localized parabolic estimate (3.3) with exponents q32, to the effect of

    vtεL2uloc,(Rd)q,,εeCεt(vL2uloc,(Rd)+t0(ts)d2(3q12)12vs3Lquloc,(Rd)ds)q,,εeCεt(vL2uloc,(Rd)+v3L(R+;Lquloc,(Rd))),

    hence vtεL([0,T];H1uloc,(Rd)).

    Substep 1.4. Conclusion: global well-posedness in H1uloc,(Rd).

    We argue that it suffices to prove the following localized energy estimate: for all T>0 and vH1uloc,(Rd), if vεL([0,T];H1uloc,(Rd)) is a weak (Duhamel) solution of

    (ε+i)tvε=vε+|vε|2vε,vε|t=0=v, (3.10)

    then we have for all 0tT,

    Euloc,(vtε)e12εtEuloc,(v). (3.11)

    Since by definition

    Euloc,(g)g2H1uloc,(Rd)+g4L4uloc,/2(Rd),

    we can naturally combine this energy estimate (3.11) together with the local well-posedness result of Step 3 with q=4. As the restriction q>6dd+2 reduces in that case to d<4, and as H1uloc,(Rd) embeds in L4uloc,(Rd) by the Sobolev embedding for d<4, this leads us to the claimed global well-posedness result.

    It remains to prove the energy estimate (3.11). If the solution vε of (3.10) was smooth, then, using the standard notation a,b:=(ˉab) for the scalar product in C, we could compute by Eq (3.10) and Young's inequality,

    tRdχx0(|vε|2+12|vε|4)=2Rdχx0(vε,tvε+|vε|2vε,tvε)=2Rdχx0vε+|vε|2vε,tvε2Rdχx0vε,tvε=2εRdχx0|tvε|22Rdχx0vε,tvε12εRdχx0|vε|2,

    hence, by Grönwall's inequality,

    Rdχx0(|vtε|2+12|vtε|4)e12εtRdχx0(|v|2+12|v|4). (3.12)

    This estimate can be justified in our non-smooth setting by an approximation procedure as e.g., in [5,6], and the claim (3.11) then follows from (3.1).

    Step 2. Global well-posedness in H1homL4hom(Rd×Ω).

    Let d<4. Given uH1homL4hom(Rd×Ω), we prove the existence of a unique almost sure global weak (Duhamel) solution uε in L(R+;H1homL4hom(Rd×Ω)) of Eq (1.2). We split the proof into four further substeps.

    Substep 2.1. Existence and uniqueness for realizations.

    Let >d2 be fixed. The localized energy of a realization u(,ω) is trivially bounded by

    Euloc,(u(,ω))Mu(ω):=zZdz2B(z)(|u(,ω)|2+12|u(,ω)|4).

    As u belongs to H1homL4hom(Rd×Ω), the choice >d2 ensures E[Mu]<. This implies that there is a subset Ω0Ω with maximal probability such that

    Euloc,(u(,ω))Mu(ω)<for all ωΩ0, (3.13)

    hence u(,ω)H1uloc,(Rd). Therefore, in view of Step 1, for all ωΩ0, there exists a unique weak (Duhamel) solution uε(,ω)Lloc(R+;H1uloc,(Rd)) of

    (ε+i)tuε(,ω)=uε(,ω)+|uε(,ω)|2uε(,ω),uε(,ω)|t=0=u(,ω), (3.14)

    and it satisfies the following energy estimate for all t0,

    Euloc,(utε(,ω))e12εtEuloc,(u(,ω))e12εtMu(ω). (3.15)

    Substep 2.2. Measurability.

    For all T we show that the above-defined map Ω0L([0,T];H1uloc,(Rd)):ωuε(,ω) is Bochner measurable.

    On the one hand, by Fubini's theorem, the joint measurability of u on Ω×Rd ensures that the map Ω0H1uloc,(Rd):ωu(,ω) is weakly measurable, hence also Bochner measurable by Pettis' theorem [1,Lemma 11.37] as H1uloc,(Rd) is a separable Banach space. On the other hand, arguing again as in Substep 1.3, we note that the solution operator H1uloc,(Rd)L([0,T];H1uloc,(Rd)):u(,ω)uε(,ω) is locally Lipschitz continuous. The Bochner measurability of uε follows by composition.

    Substep 2.3. Spatial homogeneity.

    For all T, we show that uε belongs to the space L([0,T];H1homL4hom(Rd×Ω)) up to modification on null sets.

    For all ωΩ0 and xRd, since by definition (Txu)(,ω)=u(+x,ω), cf. Section 2.2, the uniqueness of a weak (Duhamel) solution entails for almost all t,y,

    (Txutε)(y,ω)=utε(y+x,ω). (3.16)

    In other words, uε satisfies an "almost everywhere'' version of spatial homogeneity, and it remains to modify it on null sets to make it spatially homogeneous in the sense of Definition 2.1. By the measurability statement of Substep 2.2 and by the bound (3.15) with MuL1(Ω), we have uεL2(Ω0;L([0,T];H1uloc,(Rd))). Up to modification on null sets, we deduce uεL([0,T];L2(Ω;H1uloc,(Rd))), and we then define for all δ>0,

    Utε,δ(ω):=Bδutε(,ω).

    By definition and by (3.15), the family (Uε,δ)δ>0 is bounded in L([0,T];H1L4(Ω)). Up to an extraction, Uε,δ converges weakly to some Uε in that space as δ0. For all x, this implies that TxUε,δ converges weakly to TxUε. Now by (3.16), we find TxUtε,δ(ω)=Bδ(x)utε(,ω) for almost all t,x,ω. Passing to the limit yields TxUtε(ω)=utε(x,ω) for almost all t,x,ω, and the claim follows.

    Substep 2.4. Conclusion.

    It remains to check dissipation estimates. We start with mass dissipation. Let χR(x):=χ(1Rx). If the solution uε was smooth, Eq (1.2) would allow to compute, almost surely,

    tRdχR|uε(,ω)|2=2RdχRuε(,ω),tuε(,ω)=2RdχRuε(,ω),1εi(uε(,ω)+|uε(,ω)|2uε(,ω))=2εε2+1RdχR(|uε(,ω)|2+|uε(,ω)|4)2RdχRuε(,ω),1εiuε(,ω),

    hence, by integration, with |χR|1RχR,

    |RdχR|utε(,ω)|2RdχR|u(,ω)|2+2εε2+1t0RdχR(|uε(,ω)|2+|uε(,ω)|4)|1Rt0RdχR(|uε(,ω)|2+|uε(,ω)|2).

    Up to convolving uε with a smooth kernel and passing to the limit, this estimate is easily justified in our non-smooth setting. Now taking the expectation, using the spatial homogeneity, and letting R, the claimed mass dissipation identity follows.

    We turn to energy dissipation. Repeating the argument for (3.12), but replacing the exponential cut-off χx0 by χR, we get

    RdχR(|utε(,ω)|2+12|utε(,ω)|4)e12εR2tRdχR(|u(,ω)|2+12|u(,ω)|4).

    Now taking the expectation, using the spatial homogeneity, and letting R, the claimed energy dissipation estimate follows.

    The author thanks Bjoern Bringmann and Laure Saint-Raymond for bringing up this problem and for discussions on the topic, and acknowledges financial support from the CNRS-Momentum program.

    The author declares no conflict of interest.



    [1] C. D. Aliprantis, K. C. Border, Infinite dimensional analysis. A hitchhiker's guide, 3 Eds., Berlin: Springer, 2006.
    [2] J. Bourgain, Fourier transform restriction phenomena for certain lattice subsets and applications to nonlinear evolution equations. I. Schrödinger equations, Geom. Funct. Anal., 3 (1993), 107-156. doi: 10.1007/BF01896020
    [3] B. Dodson, A. Soffer, T. Spencer, The nonlinear Schrödinger equation on Z and R with bounded initial data: examples and conjectures, J. Stat. Phys., 180 (2020), 910-934. doi: 10.1007/s10955-020-02552-w
    [4] M. Duerinckx, C. Shirley, A new spectral analysis of stationary random Schrödinger operators, J. Math. Phys., 62 (2021), 072106. doi: 10.1063/5.0033583
    [5] J. Ginibre, G. Velo, On the global Cauchy problem for some nonlinear Schrödinger equations, Ann. Inst. H. Poincaré Anal. Non Linéaire, 1 (1984), 309-323. doi: 10.1016/S0294-1449(16)30425-5
    [6] J. Ginibre, G. Velo, The global Cauchy problem for the nonlinear Schrödinger equation revisited, Ann. Inst. H. Poincaré Anal. Non Linéaire, 2 (1985), 309-327. doi: 10.1016/S0294-1449(16)30399-7
    [7] J. Ginibre, G. Velo, Scattering theory in the energy space for a class of nonlinear Schrödinger equations, J. Math. Pure. Appl., 64 (1985), 363-401.
    [8] V. V. Jikov, S. M. Kozlov, O. A. Oleǐnik, Homogenization of differential operators and integral functionals, Berlin: Springer-Verlag, 1994.
    [9] Y. Katznelson, An introduction to harmonic analysis, 3 Eds., Cambridge: Cambridge University Press, 2004.
    [10] S. Nazarenko, Wave turbulence, Heidelberg: Springer, 2011.
    [11] T. Oh, Global existence for the defocusing nonlinear Schrödinger equations with limit periodic initial data, Commun. Pure Appl. Anal., 14 (2015), 1563-1580. doi: 10.3934/cpaa.2015.14.1563
    [12] T. Oh, On nonlinear Schrödinger equations with almost periodic initial data, SIAM J. Math. Anal., 47 (2015), 1253-1270. doi: 10.1137/140973384
    [13] G. C. Papanicolaou, S. R. S. Varadhan, Boundary value problems with rapidly oscillating random coefficients, In: Random fields, Vol. I, II (Esztergom, 1979), North-Holland, Amsterdam, 1981,835-873.
    [14] J. von Neumann, Über Einen Satz Von Herrn M. H. Stone, Ann. Math., 33 (1932), 567-573. doi: 10.2307/1968535
  • This article has been cited by:

    1. Jia Shen, Avy Soffer, Yifei Wu, Almost Sure Well-Posedness and Scattering of the 3D Cubic Nonlinear Schrödinger Equation, 2023, 397, 0010-3616, 547, 10.1007/s00220-022-04500-z
    2. Roberta Bianchini, Chiara Saffirio, Fluid instabilities, waves and non-equilibrium dynamics of interacting particles: a short overview, 2022, 5, 2640-3501, 1, 10.3934/mine.2023033
  • Reader Comments
  • © 2022 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(2152) PDF downloads(128) Cited by(2)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog