Global semigroup of conservative solutions of the nonlinear variational wave equation (Q715318)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Global semigroup of conservative solutions of the nonlinear variational wave equation |
scientific article |
Statements
Global semigroup of conservative solutions of the nonlinear variational wave equation (English)
0 references
5 November 2012
0 references
This paper constructs a global unique weak solution of the Cauchy problem for a model for liquid crystal evolution introduced by R. A. Saxton in 1989: \[ u_{tt}-c(u)[c(u)u_x]_x=0, \] on the real line, where \(c\in C^{1}(\mathbf R)\) takes its values in some interval \([1/\kappa,\kappa]\), \(\kappa\) being a positive constant. The main issue is to allow solutions that display a concentration of the ``energy'' \({1\over 2}(u_t^2+c^2u_x^2)\) on sets of measure zero, while retaining the conservation in time of the integral of this quantity over the whole line. This is achieved in two steps. First, the set of admissible initial data is taken to consist of three \(L^2\) functions \(u\), \(R\) and \(S\) together with two nonnegative Radon measures \(\mu\) and \(\nu\); they satisfy \(R=u_t+cu_x\), \(S=u_t-cu_x\), \(\mu_{\text{ac}}={1\over 2}R^2\) and \(\nu_{\text{ac}}={1\over 2}S^2\). The singular parts of these measures account for the concentration of ``energy.'' The measures evolve according to equations \((\mu+\nu)_t=[c(u)(\mu-\nu)]_x\) and \([{1\over c(u)}(\mu-\nu)]_t=(\mu+\nu)_x\). The second step is to cast this problem into an equivalent system \(Z_{XY}=F(Z)(Z_X,Z_Y)\), where \(X\) and \(Y\) are characteristic variables, \(F(Z)\) is bilinear and symmetric, and \(Z\) has five components; the first two components are the original independent variables \(t\) and \(x\), the third is \(u\) and the last two are ``energy potentials''. The method is, as the authors explain, closely related to the work of \textit{A. Bressan} and \textit{Y. Zheng} [Commun. Math. Phys. 266, No. 2, 471--497 (2006; Zbl 1117.35047)]. Concentration occurs, for almost all \(t\), within the set where \(c'(u(x,t))=0\). A numerical method leading to the solutions of the paper is briefly described and illustrated; it differs from the usual methods that involve numerical dissipation.
0 references
concentration phenomena
0 references
one space dimension
0 references
measure data
0 references