A note on breaking of symmetry for a class of variational problems (Q2275206): Difference between revisions

This paper considers a class of \(2\pi\)-periodic problems of the form \[ \left\{ \begin{array}{l} -u''(t) = g(u(t)) + f(t) \\ u(0) - u(2\pi) = u'(0) - u'(2\pi) = 0 \end{array} \right. \tag{\(*\)} \] where \(g: \mathbb{R}\to \mathbb{R}\) is a given \(C^1\) function and \(f:[0, 2\pi]\to \mathbb{R}\) is continuous. Two natural questions arise for this problem: \begin{itemize} \item[(1)] Which hypotheses on the nonlinearity \(g\) guarantee that, for any given \(f\) which is periodic with \textit{period} (say) \(T = 2 \pi/5\), every solution of (\(*\)) is also periodic with \textit{period} \(T = 2\pi/5\)? This is a question on \textit{Preservation of Symmetry}. \item[(2)] Which hypotheses on the nonlinearity \(g\) guarantee that there exist some periodic function \(\hat{f}\) with \textit{period} (say) \(\hat{T} = 2\pi/5\) and such that problem (\(*\)) has a periodic solution with \textit{period} (say) \(T = 2\pi/13\), different from \(\hat{T} = 2\pi/5\)? In such a case, the question is one on \textit{Breaking of Symmetry}. \end{itemize} This paper gives an answer to the Question (2) with \(\hat{T} = 2\pi/\hat{p}\) and \(T = 2\pi/p\), where \(p\) and \(\hat{p}\) are two distinct prime numbers.