Chapter 7 - On the Global Weak Solutions to a Variational Wave Equation

https://doi.org/10.1016/S1874-5717(06)80010-1Get rights and content

Abstract

A nonlinear wave equation arises in a simplified liquid crystal model through the variational principle. The wave speed of the wave equation is a given function of the wave amplitude. In the earlier study to this equation, Hunter and Saxton have derived a simple asymptotic equation for weakly nonlinear unidirectional waves of the equation. Previous work has established the existence of weak solutions to the initial value problem for the asymptotic equation for data in the space of bounded variations. We improve the previous work to the natural space of square integrable functions, and we establish the uniqueness of weak solutions for both the dissipative and conservative types.

We also have results on the full nonlinear wave equation. It has been known from joint work of the second author with Glassey and Hunter for the equation that smooth initial data may develop singularities in finite time, a sequence of weak solutions may develop concentrations, while oscillations may persist. For monotone wave speed functions in the equation, we find an invariant region in the phase space in which we discover: (a) smooth data evolve smoothly forever; (b) the smooth solutions obtained through data mollification and step (a) for not-as-smooth initial data yield weak solutions to the Cauchy problem of the nonlinear variational wave equation with initial data in H1()×L2().Furthermore, for initial data outside the invariant region, we can also prove the global existence of weak solution with initial Riemann invariant in L()L2().The main tool for the weak solution is the Young measure theory and related techniques.

More specifically, we will present the following results.

  • 1.

    On the asymptotic equation, we have existence and uniqueness of multiple weak solutions in the weak norm L2of the derivative ux.

  • 2.

    For the nonlinear wave equation, with monotone wave speed, we found some invariant regions and some global smooth solutions.

  • 3.

    For the nonlinear wave equation, with monotone wave speed, we prove the global existence of weak solution with initial Riemann invariant in L()L2().

Introduction

In this chapter we study the Cauchy problem for the nonlinear wave equation {uttc(u)[c(u)ux]x=0,t>0,x,(u,tu)|t=0=(u0,u1)(x).

The equation in (0.0.1) is the Euler–Lagrange equation of the least action principle δδu{(tu)2c2(u)(xu)2}dx dt=0.

The equation in (0.0.1) may be regarded as a generalization of equations for harmonic wave maps and arises in a number of different physical contexts, including nematic liquid crystals [54], long waves on a dipole chain in the continuum limit [30,31,77] and in classical field theories and general relativity [30]. Let us take, for example, nematic liquid crystals. We know that the mean orientation of the long molecules in a nematic liquid crystal is described by a director field of unit vectors, nS2, the unit sphere. Associated with the director field n, there is the well-known Oseen–Franck potential energy density W given by W(n,n)=α|n×(×n)|2+β(·n)2+γ(n·×n)2.

The positive constants α, β and γ are elastic constants of the liquid crystal. For the special case α = β = γ, the potential energy density reduces to W(n,n)=α|n|2,which is the potential energy density used in harmonic maps into the sphere S2. There are many studies on the constrained elliptic system of equations for n derived through variational principles from the potential (1.1.2), and on the parabolic flow associated with it, see [5,15,22,33,41,64] and references therein. In the regime in which inertia effects dominate viscosity, however, the propagation of the orientation waves in the director field may then be modeled by the least action principle [54], δδu{tn·tnW(n,n)}dx dt=0,n·n=1.

In the special case α = β = γ,this variational principle (1.1.3) yields the equation for harmonic wave maps from (1 + 3)-dimensional Minkowski space into the two sphere, see, for example, [9,57,58]. For planar deformations depending on a single space variable x, the director field has the special form n=cosu(x,t)ex+sinu(x,t)ey,where the dependent variable u1 measures the angle of the director field to the x-direction, and ex and ey are the coordinate vectors in the x and y directions, respectively. In this case, the variational principle (1.1.3) reduces to (1.1.1) with the wave speed c given specifically by c2(u)=αcos2u+βsin2u.

The general problem of global existence and uniqueness of solutions to the Cauchy problem of the nonlinear variational wave equation (0.0.1) is open. It has been demonstrated in [29] that (0.0.1) is rich in structural phenomena associated with weak solutions. Writing the highest derivatives of (0.0.1) in conservative form t2ux(c2xu)=cc(xu)2,we see that the strong precompactness in L2 of the derivatives {xu}of a sequence of approximate solutions is essential in establishing the existence of a global weak solution. However, the equation has the phenomenon of persistence of oscillation [20] and annihilation in which a sequence of exact solutions with bounded energy can oscillate forever so that the sequence {xu} is not precompact in L2, but the weak limit of the sequence is still a weak solution. Secondly, the equation has short-time smooth solutions that blow up in finite time. Thirdly, from the study of its asymptotic equation (see further), it is clear that a positive amount of energy of a solution can concentrate in a set of measure zero, and there are multiple choices for the continuation of the solution beyond blow-up time. To put the equation and its particulars into context of nonlinear wave equations under current research, we compare this equation with the equation t2ux[p(xu)]=0,where p(·) is a given function, considered by Lax [44]. The derivative xuremains bounded for (1.1.5), but we find that it is merely in L2 for (0.0.1). We note interestingly that solutions of (1.1.5) – with a “stronger” xu-dependent nonlinearity – are more regular than solutions of (0.0.1) – with an apparently “weaker” u-dependent nonlinearity. (This kind of behavior is well known for nonlinear parabolic partial differential equations.) In each case it appears that singularities develop to the maximum extent permitted by the existence of global weak solutions. For further research into (1.1.5) and its generalizations, we refer the reader to [43] and [50]. Another related equation is t2uc2(u)Δu=0considered by Lindblad [47], who established the global existence of smooth solutions of (1.1.6) with smooth, small, and spherically symmetric initial data in 3, where the large-time decay of solutions in high space dimensions is crucial. The multidimensional generalization of (0.0.1), t2uc(u). (c(u)u)=0, contains a lower-order term proportional to cc|u|2, which (1.1.6) lacks. This lower-order term is responsible for the blow-up in the derivatives of u. Finally, we note that (0.0.1) also looks related to the perturbed wave equation t2uΔu+f(u,u,u)=0,where f(u,u,u) satisfies an appropriate convexity condition (for example, f=up or f=a(tu)2+b|u|2) or some nullity condition. Blow-up for (1.1.8) with a convexity condition has been studied extensively, see [2,28,32,38,40,46,55,60,61] and [62] for more reference. Global existence and uniqueness of solutions to (1.1.8) with a nullity condition depend on the nullity structure and large time decay of solutions of the linear wave equation in higher dimensions (see [42] and references therein). Therefore (0.0.1) with the dependence of c(u) on u and the possibility of sign changes in c′(u) is familiar yet truly different.

Despite its simplicity, (0.0.1) is not easy. A study of its geometric optical solutions is helpful and interesting. Look for solutions of the form ψ(t,x)=u0+εu(εt,xc0t)+O(ε2),where u0 is a constant state and c0 = c(u0) > 0 is its speed, Hunter and Saxton [35] found that u(·, ·) satisfies (ut+uux)x=12ux2up to a scaling factor, assuming that c(u0)0.

If u0 is such that c(u0)0,c(u0)0, then the equation is (ut+u2ux)x=uux2.

In general, if u0 is such that c(k)(u0)=0,k=1,2,,n1, but c(n)(u0)0, the equation is (ut+unux)x=12nun1ux2.

Another form of the first asymptotic equation is tυ+x(uυ)=12υ2,ux=υ.

We shall not study the second and higher asymptotic equations in this short talk.

We shall prove the compactness of the approximate solution sequence by applying Young measure theory [63] and the ideas used by Lions [49] in the proof of the global existence of weak solutions to multidimensional compressible Navier–Stokes equation, and by Joly, Métivier and Rauch [39] in the rigorous justification of weakly nonlinear geometric optics for a semilinear wave equation (see also [69]). For the convenience of the reader, we quote the following lemma from [39] (see also [20,23]) that we use in this chapter.

Lemma 1.3.1.

Let U be an open subset of n, whose boundary has zero Lebesgue measure. Given a bounded family {uε(y)}Ls(U),s>1,N-valued functions, then there exist a subsequence {εj} and a measurable family of probability measures on N,{μy(·),yU}, such that for all continuous functions F(y, λ) with F(y,λ)=O(|λ|q)as|λ|, and q < s, there holds limεj0Uϕ(y)F(y,uεj(y)) dy=UNϕ(y)F(y,λ)dμy(λ) dyfor all ϕ(y)Lr(U) with compact support in the closure of U, where 1/r+q/s=1.Moreover, |λ|sdμy(λ)dylimεj0uεj(y)Lss.

Physically by approximating directly the Hamiltonian for Euler equations in the shallow water regime, Camassa and Holm [7] derived the following equation tux2tu+3uxu=2xux2u+ux3u,t>0,x.

Mathematically, (1.4.13) is obtained and proved to be formally integrable by Fuchssteiner and Fokas [25] as a bi-Hamiltonian generalization of Korteweg–de Vries (KDV) equation. Equation (1.4.13) has several important features that distinguish it from the well-known KDV equation. First, Camassa and Holm discovered that (1.4.13) possesses peaked solutions with a corner at their crest, which is in sharp contrast to the solitary waves for KDV. Second, physical water waves often break down, which can not be predicted by the solutions to the KDV equation.

Formally (1.4.13) is equivalent to {tu+uxu+xP=0,t>0,x,P(t,x)=e|xy|(u2+12(uy)2)(t,y)dy.

In [11,52], the authors proved the finite time break down of smooth solution to (1.4.14). In particular, McKean gives a necessary and sufficient criterion on the initial data for the finite-time formulation of singularities in a smooth solution to (1.4.14). Furthermore, McKean describes the blow-up process by showing the formation of cusps instead of shocks for compressible fluids. Motivated by the Young measure approach in Section 2 of this chapter (see [72]), Xin and the first author of this chapter (see [67]) proved the global existence of weak solution to (1.4.14) with initial data in H1().

Set υ=xu, by taking xto the first equation of (1.4.14), we get tυ+uxυ=12υ2P+u2.

Compared this equation with (1.2.10), we find that (1.2.10) can be considered as the first-order approximation to (1.4.14).

Up to now, we already proved the global existence and uniqueness of both dissipative and energy conservative solution to (1.2.10) through Young measure approach and explicit construction of the approximate solutions (see Section 2 of this chapter or [72]). Notice the relations between (1.2.10) and (1.4.14), we can step-by-step follow the method of [67] to prove the global existence of dissipative solution by vanishing viscosity.

Open Problem 1.

The global existence of energy conservative weak solution to (1.2.10) via the vanishing dispersion limit to (1.2.10).

As for the original variational wave equation (0.0.1), up to my recent result [75,76], we only proved the global existence of weak solution to (0.0.1) with monotone wave speed.

Open Problem 2.

The uniqueness of weak solution is open.

And finally we have the third problem.

Open Problem 3.

The global existence and uniqueness of weak solution to (0.0.1) is completely open.

Finally, let us outline the main contents of this chapter. In Section 2 we present some global existence and uniqueness results to the asymptotic equation. In Section 3 we prove some results on invariant region with monotone wave speed and in Section 4 we present one general global existence result to this equation. For some recent progress on this subject see [76].

Section snippets

References (63)

  • R. Camassa et al.

    An integrable shallow water equation with peaked solitons

    Phys. Rev. Lett.

    (1993)
  • D. Christodoulou et al.

    On the regularity of spherically symmetric wave maps

    Comm. Pure Appl. Math.

    (1993)
  • A. Constantin et al.

    Well-posedness, global existence, and blowup phenomena for a periodic quasi-linear hyperbolic equation

    Comm. Pure Appl. Math.

    (1988)
  • R.J. DiPerna et al.

    Ordinary differential equations, transport theory and Sobolev spaces

    Invent. Math.

    (1989)
  • R.J. DiPerna et al.

    Oscillations and concentrations in weak solutions of the incompressible fluid equations

    Comm. Math. Phys.

    (1987)
  • L.C. Evans

    Weak Convergence Methods for Nonlinear Partial Differential Equations

  • L.C. Evans et al.

    Measure Theory and Fine Properties of Functions

    (1992)
  • H. Fujita

    On the blowing up of solutions of the Cauchy problem for ut = †u+u1+α

    J. Fac. Sci. Univ. Tokyo Sect. I Math.

    (1966)
  • P. Gérard

    Microlocal Defect Measures

    Comm. Partial Differential Equations

    (1991)
  • R.T. Glassey

    Finite-time blow-up for solutions of nonlinear wave equations

    Math. Z.

    (1981)
  • R.T. Glassey et al.

    Singularities and oscillations in a nonlinear variational wave equation

  • A. Grundland et al.

    A.family of nonlinear Klein-Gordon equations and their solutions

    J. Math. Phys.

    (1992)
  • B. Hanouzet et al.

    Explosion pour des problemes hyperboliques semi-linéaires avec second membre non compatible

    C. R. Math. Acad. Sci. Paris

    (1985)
  • R. Hardt et al.

    Existence and partial regularity of static liquid crystal configurations

    Comm. Math. Phys.

    (1986)
  • J.K. Hunter et al.

    Dynamics of director fields

    SIAM J. Appl. Math.

    (1991)
  • J.K. Hunter et al.

    On a nonlinear hyperbolic variational equation, I and II

    Arch. Ration. Mech. Anal.

    (1995)
  • F. John

    Blow-up of solutions of nonlinear wave equations in three space dimensions

    Manuscripta. Math.

    (1979)
  • J.L. Joly et al.

    Focusing at a point and absorption of nonlinear oscillations

    Trans. Amer. Math. Soc.

    (1995)
  • T. Kato

    Blow-up of solutions of some nonlinear hyperbolic equations

    Comm. Pure Appl. Math.

    (1980)
  • Cited by (20)

    • The Hunter–Saxton equation with noise

      2021, Journal of Differential Equations
      Citation Excerpt :

      We will focus on the analysis of the characteristics following Dafermos [9]. Most of the material in this subsection can be found in classical papers by Hunter–Zheng [20,21], and also in [34]. The general solution to the nth approximation can be recovered by summing up these “boxes” defined on disjoint intervals at every t, see [20].

    • Transfer of energy in Camassa–Holm and related models by use of nonunique characteristics

      2017, Journal of Differential Equations
      Citation Excerpt :

      To prove our results, we develop the theory of nonunique characteristics, initiated in the context of conservation laws by C. Dafermos, [17] and applied by him to the Hunter–Saxton equation, [15,16,18]. This theory was then further developed by T. Cieślak and the author in [14] which led us to positive verification of the hypothesis of Zhang and Zheng, see [32], stating that dissipative solutions of the Hunter–Saxton equation dissipate energy at the highest possible rate. A similar question is also valid for the Camassa–Holm:

    • Conservative solutions to a one-dimensional nonlinear variational wave equation

      2015, Journal of Differential Equations
      Citation Excerpt :

      A numerical method for studying (1.4) was provided by Holden et al. [13]. Recently, Zhang and Zheng [30,31] have proved the global existence of conservative weak solutions for a one-dimensional system of variational wave equations, which is derived from (1.3) by considering the three-dimensional deformations of the director field that depend on a single space variable [2]. In [15,16], we dealt with a one-dimensional variational wave system arising from the variational principle (1.1) and constructed a global energy-conservative weak solution to its initial value problem.

    • Maximal dissipation in equations of evolution

      2012, Journal of Differential Equations
    View all citing articles on Scopus
    View full text