SOME NOTES ON THE FAST ADAPTIVE NEURAL NETWORK SOLVERS.

With this paper, we discuss some important aspects related to the iterative solution of two classes of polynomials, nonlinear systems of equations, and the adapted to them – "FAST adaptive neural solver" (FANS). The crucial issue of choosing initial approximations (separation of unattractive networks of initial data) and the possibility of minimizing CPU–time with the use of existing FANS is discussed.