Numerical experiments with a nonlinear evolution equation which exhibits blow-up (Q1196761)

The paper treats computation for (1) \(\dot{\omega}=[H\omega]\omega\) (where \(H\) is the Hilbert transform), proposed as a simplified vorticity model by \textit{P. Constantin}, \textit{P. D. Lax} and \textit{A. Majda} [Commun. Pure Appl. Math. 38, 715-724 (1985; Zbl 0615.76029)]\ with an explicit solution giving blow-up at \(t\to T_ *=2\) for \(\omega(0)=\cos{\pi x}\). The present concern is the extent to which standard schemes would provide a good indication that blow-up occurs; the paper explores the results of applying a standard ordinary differential equation solver to spectral and to pseudospectral discretizations for the operator. The exploration shows, in particular, the disappointing resolution of the pseudospectral method, which can be continued through \(t=2\) with comparatively indication of the blow-up despite exponential-order theoretical accuracy for smooth solutions, (i.e., \(t<2\)). To quote the authors, ``It is thus obvious that one has to be very cautious in using standard schemes when such solutions are involved''.