The long-time behaviour of solutions to parabolic problems on unbounded intervals: the influence of boundary conditions (Q4244017)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 1290040
Language Label Description Also known as
default for all languages
No label defined
    English
    The long-time behaviour of solutions to parabolic problems on unbounded intervals: the influence of boundary conditions
    scientific article; zbMATH DE number 1290040

      Statements

      The long-time behaviour of solutions to parabolic problems on unbounded intervals: the influence of boundary conditions (English)
      0 references
      0 references
      0 references
      15 March 2000
      0 references
      one-dimensional version of concentrated compactness lemma
      0 references
      Dirichlet boundary conditions
      0 references
      Neumann boundary conditions
      0 references
      Consider the initial-boundary value problem \[ u_t = u_{xx} + f(u) \quad \text{for } x>0,\;t>0, \qquad u(x,0)=u_0(x) \quad \text{for } x>0, \] with either the Dirichlet boundary \(u(0,t)=0\) or the Neumann boundary condition \(u_x(0,t)=0\) for \(t>0\). The initial function \(u_0\) is assumed nonnegative and sufficiently smooth with a unique maximum point (a so-called ``1-hump'' function), and certain technical assumptions are made on \(f\): \(f\) is sufficiently smooth with \(0 \leq f(s)\leq ks\) for large \(s\) and some positive number \(k\). In addition, \(f(0)=0\), \(f'(0)>0\), and the number \[ \zeta_0= \inf \{s>0: \int_0^s f(\sigma) d\sigma\} \] is positive and finite with \(f(\zeta_0) <0\). Under these hypotheses, it is well-known that there is a unique ground state, that is, a function \(w_g\) such that \(w_g''=f(w_g)\), \(w_g(x) \to 0\) as \(x\to\pm\infty\), and \(w_g\) attains its positive maximum at \(0\). The authors study these two problems in terms of a parameter \(\lambda\), assuming that \(u_0=\lambda \bar u\) for some fixed function \(\bar u\). For the Neumann problem, there is a positive number \(\lambda_N\) such that the solution of the Neumann problem converges to zero if \(\lambda<\lambda_N\) and to \(\infty\) if \(\lambda >\lambda_N\). Moreover it converges to \(w_g\) if \(\lambda=\lambda_N\). For the Dirichlet problem, the situation is somewhat more complicated. They show that there are numbers \(\lambda_D\) and \(\lambda_D'\) with \(\lambda_N < \lambda_D \leq \lambda_D'\) such that the solution converges to zero if \(\lambda<\lambda_D\) and to \(\infty\) if \(\lambda >\lambda_D'\). If \(\lambda_D \leq \lambda\leq \lambda_D'\), then there is a function \(y(t)\) with \(\lim_{t\to\infty} y(t)=\infty\) such that \(v_t\) defined by \(v_t(x) = u(x,t)-w_g(x-y(t))\) converges to zero as \(t\to \infty\). The new ingredient (compared to the well-studied case of the corresponding problem on a bounded interval) is the behavior of the solution to the Dirichlet problem when \(\lambda_D \leq\lambda\leq \lambda_D'\). The key step in the proof is a one-dimensional version of the concentrated compactness lemma due to P.-L. Lions.
      0 references

      Identifiers