Well-posedness and ill-posedness results for the Kadomtsev-Petviashvili-I equation (Q1395932)

This paper deals with the ``bad'' behaviour of the KP-I equation \[ (u_t+ uu_x+ u_{xxx})_x- u_{yy}= 0\tag{1} \] with respect to Picard iterative methods applied to the associated integral equation, for data in Sobolev spaces, which leads to some fluid of ill-posedness of the corresponding Cauchy problem: the flow map cannot be of class \(C^2\) in any Sobolev space. This contrasts the good behaviour of the KP-II equation (where the minus sign in (1) is replaced by the plus sign).