1: \begin{abstract}

2: ~Fast convergent, accurate, computationally efficient, parallelizable,

3: and robust matrix inversion and parameter estimation algorithms are required

4: in many time-critical and accuracy-critical applications

5: such as system identification, signal and image processing, network and big data analysis, machine learning and in many others.

6: \\

7: This paper introduces  new composite power series expansion

8: with optionally chosen rates

9: (which can be calculated simultaneously on parallel units with different computational capacities)

10: for further convergence rate improvement

11: of high order Newton-Schulz iteration.

12: New expansion was integrated into the Richardson iteration and resulted in significant

13: convergence rate  improvement.

14: The improvement is quantified via explicit transient models

15: for estimation errors and by simulations.

16: In addition, the recursive and computationally efficient version of the combination of

17: Richardson iteration and Newton-Schulz iteration with composite expansion is developed for

18: simultaneous calculations.

19: \\ Moreover, unified factorization is developed in this paper in the form

20: of tool-kit for power series expansion, which results in  a new family

21: of computationally efficient Newton-Schulz algorithms.

22: \end{abstract}