Persistent unstable equilibria and closed orbits of a singularly perturbed equation (Q1066341): Difference between revisions

Consider the bifurcation problem \(\dot y=g(x,y)\) with \(x\in {\mathbb{R}}^ a\) parameter, \(y\in {\mathbb{R}}\), and \(g(x,0)\equiv 0\). Suppose the equilibrium \(y=0\) is linearly stable for \(x<b\) and linearly unstable for \(x>b\). Now suppose the parameter x is itself slowly varying: \(\dot x=ef(x,y)\), \(f(x,0)>0\). Consider the solution that starts at \((\bar x,\delta)\), \(| \delta |\) small, \(\bar x<b\). It rapidly approaches the x-axis, moves to the right along the x-axis, and eventually rapidly leaves the x- axis. Denote its next intersection with the line \(y=\delta\) by \(p(\bar x,\epsilon,\delta)\). It is shown that \(\lim_{\epsilon \to 0}p(\bar x,\epsilon,\delta)=\tilde x\) where \(\int^{\tilde x}_{\bar x}\frac{\partial g}{\partial y}(x,0)/f(x,0)dx=0.\) The proof (in a slightly more general context) is standard and rigorous; other treatments of this problem use nonstandard analysis [\textit{F. Diener} and \textit{M. Diener}, C. R. Acad. Sci., Paris, Sér. I 297, 577-580 (1983; Zbl 0529.34019)] or asymptotic expansions [\textit{R. Haberman}, SIAM J. Appl. Math. 37, 69-106 (1979; Zbl 0417.34028)]. The formula is used to locate the limiting position of a closed orbit in a degenerate bifurcation problem from genetics.