Variation method solving of the inverse problems for Schrödinger-type equation (Q6085039)

In this paper, the authors establish the solvability of the variational formulation of inverse problems concerning the determination of unknown complex potentials in a nonlinear, non-stationary Schrödinger-type equation. The unknown potential is sought within the set of complex functions with measurable bounded components dependent on a spatial variable. Additional conditions for potential determination are presented in the form of final and boundary observations. The development of a computational algorithm for solving these inverse problems encounters various challenges. The primary difficulties involve the nonlinearity of the main equation, instability of the solution, and the implicit specification of the state operator. Beyond addressing the solvability of these problems, the authors derive a formula for the gradient of the quality criterion functional and establish the necessary conditions for optimality. These results serve as the foundation for iterative regularizing methods in solving these problems. It is worth noting that real inverse problems often fall into the category of ill-posed problems, constituting significant computational challenges. Utilizing variation methods in conjunction with iterative regularization provide universal algorithms for obtaining stable solutions to such problems. Numerical solutions for inverse and ill-posed problems can be acquired using various software packages, including Maple, Matlab, ANSYS, etc.