\documentstyle[11pt]{article}
 \textwidth 6.6in
 \textheight 8.0in
 \oddsidemargin   0in
 \evensidemargin  0in
 \topmargin  0.1in
 \parindent 0.5in
\begin{document}

 \thispagestyle{empty}

\centerline {\Large\bf  Two-sided iteration methods }
\centerline {\Large\bf for matrix inversion
\footnote{This work is partially supported by Contract MM 521/95 with
the Bulgarian Ministry of Education, Sciences and Technologies } }


\vspace{0.4cm}
\centerline {\Large\bf  Jumah A. Zarnan, Ivan G. Ivanov, Milko G. Petkov}
\vspace{0.4cm}


\large

In the present paper we suggest and study two-sided iteration methods for
matrix inversion of positive definite matrices, M-matrices and arbitrary
matrices.

We show in case symmetric positive definite matrix $A$ the two-sided iteration
process for calculating the inverse $A^{-1}$ enables us to obtain the maximum
eigenvalue of $A^{-1}$.

The two-sided methods besides their capacites of parallelizing have
the advantage of easier criteria of stopping and estimating the error of
successive approximations.


\end{document}