On the low regularity of the fifth order Kadomtsev-Petviashvili I equation (Q959859)

The paper is dealing with the existence, for finite times, of a solution to the initial-value problem for the following generalization of Kadomtsev-Petviashvili (KP) equation: \[ \partial_x[\partial_tu + \alpha\partial_x^3u + \beta\partial_x^5u + u\partial_xu] +\partial_y^2u = 0. \] In the cases of \(\beta > 0\) and \(\beta < 0\) it is called, respectively, the fifth-order KP-I and KP-II equation. Unlike their counterparts with \(\beta = 0\), i.e., the KP-I and KP-II equations proper (which correspond, severally. to \(\alpha < 0\) and \(\alpha > 0\)), the fifth- order KP equations \textit{do not} belong to the class of nonlinear PDEs integrable by means of the inverse scattering transform. The opposite signs of \(\beta\) in the case of the fifth-order equations, as well as the opposite signs of \(\alpha\) in the case of the ordinary KP equations (with \(\beta = 0\)), make dispersion relations for the linearized versions of these equations drastically different; in particular, resonance properties of the dispersion relations are different. A known consequence of this difference is that, in the case of the usual KP equations, it is much more difficult to rigorously prove the existence of a solution to the respective Cauchy problem for the equation of type I. The main point in the present work is that the smoothing effect of the fifth-order dispersion makes its easier to prove the existence of solutions for the fifth-order KP-I equation than it was for its ordinary (third-order) integrable counterpart. Using the so-called dyadic decomposed Strichartz estimate, written in terms of the respective Bourgain space, the work produces a rigorous prove of the well-posedness of the initial-value problem for the fifth-order KP-I equation in a class of interpolated energy spaces, to which the initial value, i.e., \(u(x,y,t=0)\) must belong. The well-posedness includes the existence of a single solution on a finite time interval, and the smoothness of the mapping of the initial-value function into the solution.