“…It has been the subject, for about half a century, of many investigations both from a theoretical point of view (see, e.g., [26,5,44,38,45,35,29,36]) and from its numerical approximation. In this regard, besides the first numerical approaches in [49,1,50,7], conservative methods have been developed by using various approaches [33,25], including Galerkin methods [53,41,52,6,34,22], finite difference schemes [2,55,42], operator splitting and exponential-type integrators [32,31], structure and energypreserving methods [51,23,39,43,37,48]. 1 Ideally, in the most favourable case where u is analytic, its n-th Fourier coefficient decays exponentially with n, whereas it decays as n −r if u ∈ C r .…”