Inverse scattering transform for the time dependent Schrödinger equation with applications to the KPI equation (Q915024): Difference between revisions

Rigorous results for the direct-inverse scattering problem of the time dependent Schrödinger equation \(i\psi_ y+\psi_{xx}=-u(x,y)\psi\) are obtained under the assumptions \(\| \hat u\|_{L_ 1(dldy)}<(2\pi)^{1/2}\), and \(\hat u\in L_ 2(| l|^{- 1}dldy)\), where \(\hat u(l,y)\) is the Fourier transform of \(u(x,y)\) in x. The inverse problem is a nonlocal Riemann-Hilbert problem, in which the nonlocal jump is a positive Fredholm operator under the assumptions on u. This positivity implies the solvability of the RH problem. The triangular factorization of Fredholm operators plays an important role here that may be compared to the matrix factorization in \(1+1\) inverse scattering problems. For instance, to obtain the desirable estimates, it is important to construct the equation for the inverse problem in terms of the triangular factors of the positive jump operator mentioned above. The Cauchy problem of the connected Kadomtsev-Petviashvili (I) equation \[ (u_ t+6uu_ x+u_{xxx})_ x=3u_{yy} \] is then solved for the Cauchy data \(u(x,y,t=0)\) satisfying the conditions mentioned above.