Generalization of the least squares method and regularising algorithms of the acquisition of stable
approximated solutions of systems of linear algebraic equations with approximated data appearing
with solving geophysical problems. 1
© Strakhov V.N., Strakhov A.V.
The United RAN Institute of Physics of the Earth, Moscow, Russia
Presented by editorial committee member Starostenko V.I.
The previous [105] paper shows a new characterization of the solutions
(pseudosolutions) x of systems of linear algebraic equations Ax=fσ+&f
obtained by the least squares method, namely (a(i), fσ-Ax)=0, 1, 2, ..., M, where
a(i) are vector-columns of the matrix A. In the present paper several
generalizations of the least squares method are given whose starting point is this generalization.
<<back |
|
