Abstract:
The influence of numerical differentiation errors on the convergence of methods of difference linearizations when the input data are inaccurate is studied, and the methods are optimized from the point of view of accuracy when the input data are inaccurate.