Linear recurrence relations in the algebra of matrices and applications (Q5940022)

For a sequence \(\{V_n\}\) of complex numbers with a linear recurrence relation of order \(r\geq 2\): \(V_{n+1}=\sum_{i=0}^{r-1}a_iV_{n-i}\) for \(n\geq r-1\), \textit{M.~Mouline} and \textit{M.~Rachidi} [Fibonacci Q. 37, No. 1, 34-38 (1999; Zbl 0937.11004)] have established the formula: \(V_n=\sum_{i=0}^{r-1}\rho(n-i,r)W_i\) where \(W_s=\sum_{i=s}^{r-1}a_{i}V_{r-1-i}\) (\(0\leq s\leq r-1\)) and \(\rho(n,r)\) is a polynomial of \(a_0,\dots,a_{r-1}\) with combinatorial coefficients. In this paper, the authors generalize this result to the matrix algebra. With the aid of the Cayley-Hamilton theorem, they derive some explicit formulas for \(A^n\) (\(n\geq r\)) and e\(^{tA}\) for every \(r\times r\) matrix \(A\), in terms of the coefficients of its characteristic polynomial and the matrices \(A^j\), where \(0\leq j\leq r-1\).