The paper is devoted to an algorithm for computing matrices $A^r$ and $(A^r -I).(A-I)^{-1}$ for a given square matrix $A$ and a real $r$. The algorithm uses the binary expansion of $r$ and has the logarithmic computational complexity with respect to $r$. The problem stems from the control theory.
@article{104283, author = {Jan Je\v zek}, title = {An efficient algorithm for computing real powers of a matrix and a related matrix function}, journal = {Applications of Mathematics}, volume = {33}, year = {1988}, pages = {22-32}, zbl = {0637.65036}, mrnumber = {0934371}, language = {en}, url = {http://dml.mathdoc.fr/item/104283} }
Ježek, Jan. An efficient algorithm for computing real powers of a matrix and a related matrix function. Applications of Mathematics, Tome 33 (1988) pp. 22-32. http://gdmltest.u-ga.fr/item/104283/
Theory of matrices, (in Russian). Moscow 1966. English translation: Chelsea, New York 1966. (1966)
Algol 60 Implementation, Academic Press 1964. Russian translation: Mir 1967. (1964) | MR 0215554
The art of computer programming, vol 2, Addison-Wesley 1969. Russian translation: Mir 1977. (1969) | MR 0633878 | Zbl 0191.18001
Computation of matrix exponential, square root and logarithm, (in Czech). Knižnica algoritmov, diel III, symposium Algoritmy, SVTS Bratislava 1975. (1975)
General matrix power and sum of matrix powers, (in Czech). Knižnica algoritmov, diel IX, symposium Algoritmy, SVTS Bratislava 1987. (1987)