Generalized Fibonacci sequences via orthogonal polynomials (Q440967)

For any four real numbers \(a, b, c\) and \(d\), the generalized Fibonacci sequence \(\{ Q_n \}^{\infty}_{n=0}\) is defined recursively by \[ Q_0 = 0, \;Q_1 = 1, \] \[ Q_n = \left \{ \begin{matrix} aQ_{n-1} + cQ_{n-2}, & \text{ if \(n\) is even} \\ bQ_{n-1} + dQ_{n-2}, & \text{ if \(n\) is odd} \end{matrix} \right. (n \geq 2). \] A new proof is given of O. Yayenie's Theorem, that is: the \(m\)th term of the generalized Fibonacci sequence \(\{ Q_n \}^{\infty}_{n=0}\) defined by the above formula, is given by \[ Q_m = \frac{a^{1-\zeta(m)}}{(ab)^{[m/2]}}.\frac{\alpha^{[m/2]}(\alpha + d - c)^{m-[m/2]} - \beta^{[m/2]}(\beta + d - c)^{m-[m/2]}}{\alpha - \beta}, \] where \(\zeta(m) = m - 2\left[ \frac{m}{2} \right ],\) and \(\alpha\) and \(\beta\) are the roots of the quadratic equation \[ x^2 - (ab+c-d)x-abd = 0. \] Fix an integer number \(k \geq 2\), and let \(a_0, a_1, ..., a_{k-1}\) and \(b_0, b_1, ..., b_{k-1}\) be \(2k\) given real or complex numbers, with \(b_j \not= 0\) for \(0 \leq j \leq k-1\). Let \(\{ Q_n \}^{\infty}_{n=0}\) be the sequence of real or complex numbers defined recursively by \[ Q_0 = 0, \;Q_1 = 1, \;Q_m = a_jQ_{m-1} + b_jQ_{m-2}, \;m \equiv j (\mod k). \] It is proved that for \(k \geq 3\), the term of the generalized Fibonacci sequence \(\{ Q_n \}^{\infty}_{n=0}\) defined by the previous formula is given by the following Binet's-type formula: \[ Q_{n+1} = \frac{(A_{k,n}\alpha + B_{k,n})\alpha^{[m/2]} - (A_{k,n}\beta + B_{k,n})\beta^{[m/2]}}{\alpha - \beta}, \;n \geq 0, \] where \(\alpha\) and \(\beta\) are the roots of the quadratic equation \[ z^2 - (c-\Delta_k)z+b = 0, \] being \[ b = (-1)^k\prod_{i=0}^{k-1} b_i, \;c = (-1)^k(b_2+b/b_2), \] \[ A_{k,n} = \Delta_{2,1+n-k[n/k]}, \;B_{k,n} = (-1)^{1+n-k[n/k]}\left ( \prod_{i=2}^{2+n-k[n/k]} b_i \right ) \Delta_{3+n-k[n/k],k}, \] where for \(0 \leq \mu < \nu \leq k\), \[ \Delta_{\mu,\nu} = \left | \begin{matrix} a_{\mu} & 1 \\ -b_{\mu+1} & a_{\mu+1} & 1 \\ & \ddots & \ddots & \ddots \\ & & -b_{\nu-1} & a_{\nu-1} & 1 \\ & & & -b_{\nu} & a_{\nu} \end{matrix} \right | \] and \[ \Delta_k = \left | \begin{matrix} a.o_2 & 1 & & & & 1 \\ -b_3 & a_3 & 1 \\ & \ddots & \ddots & \ddots \\ & & -b_{k-1} & a_{k-1} & 1 \\ -b_2& & & & -b_1 & a_1 \end{matrix} \right |. \]