Variational description of Gibbs-non-Gibbs dynamical transitions for the Curie-Weiss model (Q1950363)

From MaRDI portal
Revision as of 19:02, 21 March 2024 by Maintenance script (talk | contribs) (rollbackEdits.php mass rollback)
scientific article
Language Label Description Also known as
English
Variational description of Gibbs-non-Gibbs dynamical transitions for the Curie-Weiss model
scientific article

    Statements

    Variational description of Gibbs-non-Gibbs dynamical transitions for the Curie-Weiss model (English)
    0 references
    0 references
    13 May 2013
    0 references
    In this paper the authors study the large-deviation rate function for the trajectory of the magnetization in the Curie-Weiss model with potential \(J>0\) and magnetic field \(h\in \mathbb{R}.\) The key object of the study is the time-evolved single-spin average conditional on the final empirical magnetization. Non-Gibbsianness corresponds to a discontinuous dependence of this average on the final magnetization. The discontinuity points are called bad magnetization. Such discontinuities may arise whenever there is more than one possible trajectory compatible with the bad magnetization at the end. The main results of the paper are: {\parindent=6mm \begin{itemize}\item[1.] If \(0<J\leq 1\), then the evolved state is Gibbs at all times. If \(J>1\) there exists some time \(\psi_U\) at which multiple trajectories appear. The associated non-Gibbsianness persists for all later times when \(h=0\). \item[2.] For \(h\neq 0\) there is a time \(\psi_*>\psi_U\) at which Gibbsianness is restored for all latter times. \item[3.] There is a change in behavior at \(J=\frac{3}{2}\). For \(1<J\leq \frac{3}{2}:\) {\parindent=12mm \begin{itemize}\item[(a)] If \(h=0\) there exists some time \(\psi_c\) such that only the zero magnetization is bad for \(t>\psi_c\). \item[(b)] If \(h>0 \, (h<0)\), then there is only one bad magnetization for \(\psi_U<t\leq \psi_*\). This bad magnetization changes with \(t\) but is always strictly negative (positive). \end{itemize}} \item[4.] For \(J> \frac{3}{2}:\) {\parindent=12mm \begin{itemize}\item[(a)] If \(h=0\), then there is a time \(\psi_c>\psi_U\) such that for \(\psi_U<t<\psi_c\) there are two non-zero bad magnetizations, while for \(t\geq \psi_c\) only the zero magnetization is bad. \item[(b)] If \(h\neq 0\) and small enough, then there are two times \(\psi_T>\psi_L\) between \(\psi_U\) and \(\psi_*\) such that for \(\psi_U<t\leq \psi_L\) and \(\psi_T\leq t\leq \psi_*\) only one bad magnetization occurs, while for \(\psi_L<t<\psi_T\) two bad magnetizations occur. \end{itemize}} \end{itemize}}
    0 references
    0 references
    Curie-Weiss model
    0 references
    Gibbs measure
    0 references
    non-Gibbs measure
    0 references
    time
    0 references
    magnetization
    0 references