r-Fibonacci and r-Lucas polynomials

DefinitionThe \(r\)-Fibonacci polynomials#✓fully proved

Let \(R\) be a commutative semiring, let \(r\) be a natural number and let \(x, y \in R\). The \(r\)-Fibonacci polynomials \(U^{(r)}_n(x, y) \in R\), \(n \ge 0\), are defined by \(U^{(r)}_0 = 0\), \(U^{(r)}_k = x^{k-1}\) for \(1 \le k \le r\), and \(U^{(r)}_{n+1} = x \, U^{(r)}_n + y \, U^{(r)}_{n-r}\) for \(n \ge r\).

These are the initial values and recurrence displayed for \(\xi_n = m^{n-1} U_n\) in the paper, at \(m = 1\).

public def rFibonacciPolynomial {R : Type*} [CommSemiring R] (r : ℕ) (x y : R) : ℕ → R
  | 0 => 0
  | n + 1 =>
    if n < r then x ^ n
    else x * rFibonacciPolynomial r x y n + y * rFibonacciPolynomial r x y (n - r)
termination_by n => n
decreasing_by all_goals omega
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 4.
Used by 9

TheoremThe zeroth \(r\)-Fibonacci polynomial vanishes#✓fully proved

For every natural number \(r\) and all \(x, y\) in a commutative semiring \(R\), the \(r\)-Fibonacci polynomial \(U^{(r)}_0(x, y)\) is \(0\).

public theorem rFibonacciPolynomial_zero {R : Type*} [CommSemiring R] (r : ℕ) (x y : R) :
    rFibonacciPolynomial r x y 0 = 0
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 4.
Uses 1 · Used by 1

TheoremInitial values of the \(r\)-Fibonacci polynomials#✓fully proved

Let \(R\) be a commutative semiring, let \(x, y \in R\), and let \(n < r\) be natural numbers. Then the \(r\)-Fibonacci polynomial \(U^{(r)}_{n+1}(x, y)\) equals \(x^n\).

public theorem rFibonacciPolynomial_succ_of_lt {R : Type*} [CommSemiring R] {r n : ℕ}
    (x y : R) (h : n < r) : rFibonacciPolynomial r x y (n + 1) = x ^ n
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 4.
Uses 1 · Used by 1

TheoremRecurrence for the \(r\)-Fibonacci polynomials#✓fully proved

Let \(R\) be a commutative semiring, let \(x, y \in R\), and let \(r \le n\) be natural numbers. Then the \(r\)-Fibonacci polynomials \(U_m = U^{(r)}_m(x, y)\) satisfy \(U_{n+1} = x \, U_n + y \, U_{n-r}\).

public theorem rFibonacciPolynomial_succ_of_le {R : Type*} [CommSemiring R] {r n : ℕ}
    (x y : R) (h : r ≤ n) :
    rFibonacciPolynomial r x y (n + 1)
      = x * rFibonacciPolynomial r x y n + y * rFibonacciPolynomial r x y (n - r)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 4.
Uses 1 · Used by 1

TheoremThe \(1\)-Fibonacci polynomials at \(x = y = 1\) are the Fibonacci numbers#✓fully proved

For every natural number \(n\), the \(r\)-Fibonacci polynomial with \(r = 1\), evaluated at \(x = y = 1\) in the natural numbers, is the \(n\)-th Fibonacci number: \(U^{(1)}_n(1, 1) = F_n\), where \(F_0 = 0\), \(F_1 = 1\) and \(F_{n+2} = F_{n+1} + F_n\).

public theorem rFibonacciPolynomial_one_one_one :
    ∀ n, rFibonacciPolynomial 1 (1 : ℕ) 1 n = Nat.fib n
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 1.
Uses 4

TheoremGeometrically weighted sum of \(r\)-Fibonacci polynomials#●ready to prove

Let \(K\) be a field, let \(r \ge 1\) be a natural number, let \(x, y, m \in K\) with \(1 - m x - m^{r+1} y \ne 0\), and write \(U_j\) for the \(r\)-Fibonacci polynomial \(U^{(r)}_j(x, y)\). Then for every natural number \(n\),

\[ \sum_{j=1}^{n} m^{j-1} U_j = \frac{1 - m^n U_{n+1} - m^{r+1} y \sum_{j=1}^{r} m^{n-r+j-1} U_{n-r+j}}{1 - m x - m^{r+1} y}, \]

where the terms of the inner sum with \(n - r + j \le 0\) are zero.

The paper takes \(m\) a positive integer; here \(m\) is any element of \(K\).

public theorem_wanted sum_pow_mul_rFibonacciPolynomial {K : Type*} [Field K] {r : ℕ}
    (hr : 1 ≤ r) (x y m : K) (hP : 1 - m * x - m ^ (r + 1) * y ≠ 0) (n : ℕ) :
    ∑ j ∈ Finset.Icc 1 n, m ^ (j - 1) * rFibonacciPolynomial r x y j
      = (1 - m ^ n * rFibonacciPolynomial r x y (n + 1)
          - m ^ (r + 1) * y * ∑ j ∈ Finset.Icc 1 r,
              m ^ (n + j - r - 1) * rFibonacciPolynomial r x y (n + j - r))
        / (1 - m * x - m ^ (r + 1) * y)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 4.
Uses 1

TheoremSum of the first \(n\) \(r\)-Fibonacci polynomials#●ready to prove

Let \(K\) be a field, let \(r \ge 1\) be a natural number, let \(x, y \in K\) with \(1 - x - y \ne 0\), and write \(U_j\) for the \(r\)-Fibonacci polynomial \(U^{(r)}_j(x, y)\). Then for every natural number \(n\),

\[ \sum_{j=1}^{n} U_j = \frac{1 - U_{n+1} - y \sum_{j=1}^{r} U_{n-r+j}}{1 - x - y}, \]

where the terms of the inner sum with \(n - r + j \le 0\) are zero.

public theorem_wanted sum_rFibonacciPolynomial {K : Type*} [Field K] {r : ℕ} (hr : 1 ≤ r)
    (x y : K) (hP : 1 - x - y ≠ 0) (n : ℕ) :
    ∑ j ∈ Finset.Icc 1 n, rFibonacciPolynomial r x y j
      = (1 - rFibonacciPolynomial r x y (n + 1)
          - y * ∑ j ∈ Finset.Icc 1 r, rFibonacciPolynomial r x y (n + j - r))
        / (1 - x - y)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Corollary 2.
Uses 1

TheoremAlternating sum of \(r\)-Fibonacci polynomials#●ready to prove

Let \(K\) be a field, let \(r \ge 1\) be a natural number, let \(x, y \in K\) with \(1 + x + (-1)^r y \ne 0\), and write \(U_j\) for the \(r\)-Fibonacci polynomial \(U^{(r)}_j(x, y)\). Then for every natural number \(n\),

\[ \sum_{j=1}^{n} (-1)^{j-1} U_j = \frac{1 - (-1)^n U_{n+1} + (-1)^r y \sum_{j=1}^{r} (-1)^{n-r+j-1} U_{n-r+j}}{1 + x + (-1)^r y}, \]

where the terms of the inner sum with \(n - r + j \le 0\) are zero.

public theorem_wanted sum_neg_one_pow_mul_rFibonacciPolynomial {K : Type*} [Field K] {r : ℕ}
    (hr : 1 ≤ r) (x y : K) (hP : 1 + x + (-1) ^ r * y ≠ 0) (n : ℕ) :
    ∑ j ∈ Finset.Icc 1 n, (-1) ^ (j - 1) * rFibonacciPolynomial r x y j
      = (1 - (-1) ^ n * rFibonacciPolynomial r x y (n + 1)
          + (-1) ^ r * y * ∑ j ∈ Finset.Icc 1 r,
              (-1) ^ (n + j - r - 1) * rFibonacciPolynomial r x y (n + j - r))
        / (1 + x + (-1) ^ r * y)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Corollary 3.
Uses 1

DefinitionThe \(r\)-Lucas polynomials of type \(s\)#✓fully proved

Let \(R\) be a commutative semiring, let \(r\) and \(s\) be natural numbers and let \(x, y \in R\). The \(r\)-Lucas polynomials of type \(s\) are the sequence \(V_n = V^{(r,s)}_n(x, y)\), \(n \ge 0\), of elements of \(R\) given by \(V_0 = s + 1\), \(V_k = x^k\) for \(1 \le k \le r\), and \(V_{n+1} = x V_n + y V_{n-r}\) for \(n \ge r\).

The paper takes \(1 \le s \le r\); the definition makes sense for every natural number \(s\).

public def rLucasPolynomial {R : Type*} [CommSemiring R] (r s : ℕ) (x y : R) : ℕ → R
  | 0 => (s : R) + 1
  | n + 1 =>
    if n < r then x ^ (n + 1)
    else x * rLucasPolynomial r s x y n + y * rLucasPolynomial r s x y (n - r)
termination_by n => n
decreasing_by all_goals omega
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 1.
Used by 8

TheoremInitial value of the \(r\)-Lucas polynomials#✓fully proved

For every commutative semiring \(R\), all natural numbers \(r\) and \(s\) and all \(x, y \in R\), the \(r\)-Lucas polynomial of type \(s\) satisfies \(V^{(r,s)}_0(x, y) = s + 1\).

public theorem rLucasPolynomial_zero {R : Type*} [CommSemiring R] (r s : ℕ) (x y : R) :
    rLucasPolynomial r s x y 0 = (s : R) + 1
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 1.
Uses 1

TheoremInitial values of the \(r\)-Lucas polynomials#✓fully proved

Let \(R\) be a commutative semiring, let \(s\) be a natural number, let \(x, y \in R\), and let \(n < r\) be natural numbers. Then the \(r\)-Lucas polynomial of type \(s\) satisfies \(V^{(r,s)}_{n+1}(x, y) = x^{n+1}\).

public theorem rLucasPolynomial_succ_of_lt {R : Type*} [CommSemiring R] {r n : ℕ} (s : ℕ)
    (x y : R) (h : n < r) : rLucasPolynomial r s x y (n + 1) = x ^ (n + 1)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 1.
Uses 1

TheoremRecurrence for the \(r\)-Lucas polynomials#✓fully proved

Let \(R\) be a commutative semiring, let \(s\) be a natural number, let \(x, y \in R\), and let \(r \le n\) be natural numbers. Then the \(r\)-Lucas polynomials of type \(s\), \(V_m = V^{(r,s)}_m(x, y)\), satisfy \(V_{n+1} = x \, V_n + y \, V_{n-r}\).

public theorem rLucasPolynomial_succ_of_le {R : Type*} [CommSemiring R] {r n : ℕ} (s : ℕ)
    (x y : R) (h : r ≤ n) :
    rLucasPolynomial r s x y (n + 1)
      = x * rLucasPolynomial r s x y n + y * rLucasPolynomial r s x y (n - r)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 1.
Uses 1

TheoremBinet formula for the \(r\)-Fibonacci polynomials (distinct zeros)#●ready to prove

Let \(K\) be a field, let \(r \ge 1\) be a natural number and let \(x, y \in K\). Suppose that \(t^{r+1} - x t^r - y = \prod_{k=0}^{r} (t - \alpha_k)\) in \(K[t]\), where \(\alpha_0, \dots, \alpha_r \in K\) are pairwise distinct. Then for every \(n \ge 0\) the \(r\)-Fibonacci polynomials \(U^{(r)}_n(x, y)\) satisfy

\[U^{(r)}_{n+1}(x, y) = \sum_{k=0}^{r} \frac{\alpha_k^{n+1}}{(r+1) \alpha_k - r x}.\]

The paper's hypothesis \(y \ne -\frac{1}{r} \left( \frac{rx}{r+1} \right)^{r+1}\) is replaced by the distinctness of the zeros it is meant to guarantee (it does not exclude the multiple zero \(0\) when \(y = 0\), \(r \ge 2\)).

public theorem_wanted rFibonacciPolynomial_eq_sum_roots {K : Type*} [Field K] {r : ℕ}
    (hr : 1 ≤ r) (x y : K) (α : Fin (r + 1) → K) (hα : Function.Injective α)
    (hP : (Polynomial.X ^ (r + 1) - Polynomial.C x * Polynomial.X ^ r - Polynomial.C y
      : Polynomial K) = ∏ k, (Polynomial.X - Polynomial.C (α k))) (n : ℕ) :
    rFibonacciPolynomial r x y (n + 1)
      = ∑ k, α k ^ (n + 1) / ((r + 1) * α k - r * x)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 1.
Uses 1

TheoremBinet formula for the \(r\)-Lucas polynomials (distinct zeros)#●ready to prove

Let \(K\) be a field, let \(r \ge 1\) and \(s\) be natural numbers, and let \(x, y \in K\). Suppose that \(t^{r+1} - x t^r - y = \prod_{k=0}^{r} (t - \alpha_k)\) in \(K[t]\), where \(\alpha_0, \dots, \alpha_r \in K\) are pairwise distinct. Then for every \(n \ge 0\) the \(r\)-Lucas polynomials of type \(s\) satisfy

\[V^{(r,s)}_n(x, y) = \sum_{k=0}^{r} \frac{\alpha_k^n ((s+1)\alpha_k - s x)}{(r+1)\alpha_k - r x}.\]

The paper takes \(1 \le s \le r\); the identity holds for every \(s\).

public theorem_wanted rLucasPolynomial_eq_sum_roots {K : Type*} [Field K] {r : ℕ}
    (hr : 1 ≤ r) (s : ℕ) (x y : K) (α : Fin (r + 1) → K) (hα : Function.Injective α)
    (hP : (Polynomial.X ^ (r + 1) - Polynomial.C x * Polynomial.X ^ r - Polynomial.C y
      : Polynomial K) = ∏ k, (Polynomial.X - Polynomial.C (α k))) (n : ℕ) :
    rLucasPolynomial r s x y n
      = ∑ k, α k ^ n * ((s + 1) * α k - s * x) / ((r + 1) * α k - r * x)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 1.
Uses 1

TheoremBinet formula for the r-Fibonacci polynomials with one double zero#●ready to prove

Let \(K\) be a field of characteristic zero, let \(r \ge 1\) be a natural number, and let \(x, y, \alpha \in K\). Let \(\beta_1, \dots, \beta_{r-1} \in K\) be pairwise distinct and all different from \(\alpha\), put \(\tilde P(t) = \prod_{k=1}^{r-1} (t - \beta_k)\), and suppose that \(t^{r+1} - x t^r - y = (t - \alpha)^2 \tilde P(t)\) as polynomials over \(K\). Then for every natural number \(n\) the \(r\)-Fibonacci polynomial \(U^{(r)}\) satisfies

\[ U^{(r)}_{n+1}(x, y) = \frac{\alpha^{n+r-1}}{\tilde P(\alpha)} \Bigl((n + r) - \frac{\alpha \tilde P'(\alpha)}{\tilde P(\alpha)}\Bigr) + \sum_{k=1}^{r-1} \frac{\beta_k^{n+1}}{(r+1)\beta_k - r x}. \]

The paper's hypothesis \(y = -\frac{1}{r}\bigl(\frac{rx}{r+1}\bigr)^{r+1}\) (which makes \(\alpha = rx/(r+1)\) the double zero) is replaced by the factorization it yields.

public theorem_wanted rFibonacciPolynomial_eq_of_double_root {K : Type*} [Field K]
    [CharZero K] {r : ℕ} (hr : 1 ≤ r) (x y α : K) (β : Fin (r - 1) → K)
    (hβ : Function.Injective β) (hβα : ∀ k, β k ≠ α)
    (hP : (Polynomial.X ^ (r + 1) - Polynomial.C x * Polynomial.X ^ r - Polynomial.C y
      : Polynomial K) = (Polynomial.X - Polynomial.C α) ^ 2 *
        ∏ k, (Polynomial.X - Polynomial.C (β k))) (n : ℕ) :
    rFibonacciPolynomial r x y (n + 1)
      = α ^ (n + r - 1) / (∏ k, (Polynomial.X - Polynomial.C (β k))).eval α *
          ((n + r : K) - α * (Polynomial.derivative
              (∏ k, (Polynomial.X - Polynomial.C (β k)))).eval α /
            (∏ k, (Polynomial.X - Polynomial.C (β k))).eval α)
        + ∑ k, β k ^ (n + 1) / ((r + 1) * β k - r * x)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 1.
Uses 1

TheoremBinet formula for the \(r\)-Lucas polynomials (one double zero)#●ready to prove

Let \(K\) be a field of characteristic zero, let \(r \ge 1\) and \(s\) be natural numbers, and let \(x, y, \alpha \in K\) with \(\alpha \ne 0\). Suppose that \(t^{r+1} - x t^r - y = (t - \alpha)^2 \tilde P(t)\) in \(K[t]\), where \(\tilde P(t) = \prod_{k=1}^{r-1} (t - \beta_k)\) for pairwise distinct \(\beta_1, \dots, \beta_{r-1} \in K\), all different from \(\alpha\). Then for every \(n \ge 0\) the \(r\)-Lucas polynomials of type \(s\) satisfy

\[V^{(r,s)}_n(x, y) = \frac{\alpha^{n+r-1}}{\tilde P(\alpha)} \left( n \, \frac{r-s}{r} + r + \frac{s}{r} - \alpha \, \frac{r-s}{r} \, \frac{\tilde P'(\alpha)}{\tilde P(\alpha)} \right) + \sum_{k=1}^{r-1} \frac{\beta_k^n ((s+1) \beta_k - s x)}{(r+1) \beta_k - r x}.\]

The paper's hypothesis \(y = -\frac{1}{r} \left( \frac{rx}{r+1} \right)^{r+1}\) (which makes \(\alpha = rx/(r+1)\) the double zero) is replaced by the factorization it yields. The paper takes \(1 \le s \le r\); the identity holds for every \(s\). The hypothesis \(\alpha \ne 0\) is added: for \(r = 2\), \(y = 0\) the factorization \(t^2 (t - x)\) has \(\alpha = 0\) and the formula fails at \(n = 0\) (for \(x = s = 1\): \(V_0 = 2\), right side \(1\)). In the paper's setting \(\alpha = rx/(r+1)\) with \(x \ne 0\), so \(\alpha \ne 0\); the only other \(\alpha = 0\) case, \(r = 1\), \(x = y = 0\), is dropped.

public theorem_wanted rLucasPolynomial_eq_of_double_root {K : Type*} [Field K]
    [CharZero K] {r : ℕ} (hr : 1 ≤ r) (s : ℕ) (x y α : K) (hα : α ≠ 0) (β : Fin (r - 1) → K)
    (hβ : Function.Injective β) (hβα : ∀ k, β k ≠ α)
    (hP : (Polynomial.X ^ (r + 1) - Polynomial.C x * Polynomial.X ^ r - Polynomial.C y
      : Polynomial K) = (Polynomial.X - Polynomial.C α) ^ 2 *
        ∏ k, (Polynomial.X - Polynomial.C (β k))) (n : ℕ) :
    rLucasPolynomial r s x y n
      = α ^ (n + r - 1) / (∏ k, (Polynomial.X - Polynomial.C (β k))).eval α *
          (n * ((r - s) / r) + (r + s / r) - α * ((r - s) / r) *
            (Polynomial.derivative (∏ k, (Polynomial.X - Polynomial.C (β k)))).eval α /
              (∏ k, (Polynomial.X - Polynomial.C (β k))).eval α)
        + ∑ k, β k ^ n * ((s + 1) * β k - s * x) / ((r + 1) * β k - r * x)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 1.
Uses 1

Definition\((p,q)\)-binomial coefficient#✓fully proved

Let \(R\) be a commutative semiring and \(p, q \in R\). For natural numbers \(n\) and \(k\), the \((p,q)\)-binomial coefficient \(\binom{n}{k}_{p,q} \in R\) is defined recursively by \(\binom{n}{0}_{p,q} = 1\), \(\binom{0}{k+1}_{p,q} = 0\) and \(\binom{n+1}{k+1}_{p,q} = p^{n-k} \binom{n}{k}_{p,q} + q^{k+1} \binom{n}{k+1}_{p,q}\), where \(n - k\) is truncated subtraction (\(0\) when \(k > n\)).

For \(k \le n\), where the \((p,q)\)-factorials are invertible, \(\binom{n}{k}_{p,q}\) equals \([n]_{p,q}! / ([k]_{p,q}! \, [n-k]_{p,q}!)\), where \([j]_{p,q} = p^{j-1} + p^{j-2} q + \dots + q^{j-1}\); this is the definition used in the source. At \(p = 1\) it is the Gaussian binomial coefficient \(\binom{n}{k}_q\).

public def pqBinomial {R : Type*} [CommSemiring R] (p q : R) : ℕ → ℕ → R
  | _, 0 => 1
  | 0, _ + 1 => 0
  | n + 1, k + 1 => p ^ (n - k) * pqBinomial p q n k + q ^ (k + 1) * pqBinomial p q n (k + 1)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 3, Equation (3).
Used by 8

Theorem\((p,q)\)-binomial coefficients with lower index zero#✓fully proved

For all elements \(p, q\) of a commutative semiring and every natural number \(n\), the \((p,q)\)-binomial coefficient \(\binom{n}{0}_{p,q}\) equals \(1\).

public theorem pqBinomial_zero_right {R : Type*} [CommSemiring R] (p q : R) (n : ℕ) :
    pqBinomial p q n 0 = 1
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 3.
Uses 1 · Used by 2

Theorem\((p,q)\)-binomial coefficients with upper index zero and positive lower index#✓fully proved

For all elements \(p, q\) of a commutative semiring and every natural number \(k\), the \((p,q)\)-binomial coefficient \(\binom{0}{k+1}_{p,q}\) equals \(0\).

public theorem pqBinomial_zero_succ {R : Type*} [CommSemiring R] (p q : R) (k : ℕ) :
    pqBinomial p q 0 (k + 1) = 0
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 3.
Uses 1

TheoremPascal recursion for \((p,q)\)-binomial coefficients#✓fully proved

For all elements \(p, q\) of a commutative semiring and all natural numbers \(n\) and \(k\), the \((p,q)\)-binomial coefficients satisfy \(\binom{n+1}{k+1}_{p,q} = p^{n-k} \binom{n}{k}_{p,q} + q^{k+1} \binom{n}{k+1}_{p,q}\), where \(n - k\) is truncated subtraction (\(0\) when \(k > n\)).

public theorem pqBinomial_succ_succ {R : Type*} [CommSemiring R] (p q : R) (n k : ℕ) :
    pqBinomial p q (n + 1) (k + 1)
      = p ^ (n - k) * pqBinomial p q n k + q ^ (k + 1) * pqBinomial p q n (k + 1)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 3, Equation (3).
Uses 1 · Used by 1

TheoremThe \((p, q)\)-binomial coefficients at \(p = 1\)#✓fully proved

Let \(R\) be a commutative semiring and \(q \in R\). For all natural numbers \(n\) and \(k\), the \((p, q)\)-binomial coefficient \([n, k]_{p,q}\) at \(p = 1\) equals the Gaussian binomial coefficient \(\binom{n}{k}_q\).

public theorem pqBinomial_one_left {R : Type*} [CommSemiring R] (q : R) (n k : ℕ) :
    pqBinomial 1 q n k = MathlibExt.gaussBinom q n k
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Section 3 (Remark 1: p = 1 gives the q-analogue).
Uses 8

Definition\((p,q)\)-analogue of the \(r\)-Fibonacci polynomial#✓fully proved

Let \(R\) be a commutative semiring, let \(r\) and \(m\) be natural numbers and let \(x, y, p, q \in R\). The \((p,q)\)-analogue \(U^{(r)}_n(x, y, p, q, m) \in R\) of the \(r\)-Fibonacci polynomial is defined by \(U^{(r)}_0(x, y, p, q, m) = 0\) and, for every natural number \(n\),

\[ U^{(r)}_{n+1}(x, y, p, q, m) = \sum_{k=0}^{\lfloor n/(r+1) \rfloor} p^{\binom{n+1-(r+1)k}{2}} q^{\binom{k+1}{2} + m \binom{k}{2}} \binom{n-rk}{k}_{p,q} x^{n-(r+1)k} y^k, \]

where \(\binom{a}{b}_{p,q}\) is the \((p,q)\)-binomial coefficient.

The source also states \(U^{(r)}_j = p^{\binom{j+1}{2}} x^{j-1}\) for \(1 \le j \le r\), which contradicts this sum: the sum gives \(p^{\binom{j}{2}} x^{j-1}\) (for example \(U^{(2)}_2 = p x\), not \(p^3 x\)). The sum, which Theorem 2 of the source requires, is used.

public def pqRFibonacci {R : Type*} [CommSemiring R] (r : ℕ) (x y p q : R) (m : ℕ) : ℕ → R
  | 0 => 0
  | n + 1 => ∑ k ∈ Finset.range (n / (r + 1) + 1),
      p ^ (n + 1 - (r + 1) * k).choose 2 * q ^ ((k + 1).choose 2 + m * k.choose 2) *
        pqBinomial p q (n - r * k) k * x ^ (n - (r + 1) * k) * y ^ k
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 1.
Uses 1 · Used by 10

TheoremZeroth value of the \((p,q)\)-analogue of the \(r\)-Fibonacci polynomial#✓fully proved

For all natural numbers \(r\) and \(m\) and all elements \(x, y, p, q\) of a commutative semiring, the \((p,q)\)-analogue of the \(r\)-Fibonacci polynomial satisfies \(U^{(r)}_0(x, y, p, q, m) = 0\).

public theorem pqRFibonacci_zero {R : Type*} [CommSemiring R] (r : ℕ) (x y p q : R)
    (m : ℕ) : pqRFibonacci r x y p q m 0 = 0
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 1.
Uses 1

TheoremFirst value of the \((p,q)\)-analogue of the \(r\)-Fibonacci polynomial#✓fully proved

For all natural numbers \(r\) and \(m\) and all elements \(x, y, p, q\) of a commutative semiring, the \((p,q)\)-analogue of the \(r\)-Fibonacci polynomial satisfies \(U^{(r)}_1(x, y, p, q, m) = 1\).

public theorem pqRFibonacci_one {R : Type*} [CommSemiring R] (r : ℕ) (x y p q : R)
    (m : ℕ) : pqRFibonacci r x y p q m 1 = 1
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 1.
Uses 3

TheoremFirst recurrence for the \((p,q)\)-analogue of the \(r\)-Fibonacci polynomial#●ready to prove

Let \(R\) be a commutative semiring, let \(r \ge 1\), \(n \ge r\) and \(m\) be natural numbers and let \(x, y, p, q \in R\). Write \(U_j(x, y)\) for \(U^{(r)}_j(x, y, p, q, m)\), the \((p,q)\)-analogue of the \(r\)-Fibonacci polynomial with \(r\), \(p\), \(q\) and \(m\) fixed. Then

\[U_{n+1}(x, y) = p x \, U_n(p x, q y) + q y \, U_{n-r}(p x, q^{m+1} y).\]
public theorem_wanted pqRFibonacci_succ_eq_left {R : Type*} [CommSemiring R] {r n : ℕ}
    (hr : 1 ≤ r) (hn : r ≤ n) (x y p q : R) (m : ℕ) :
    pqRFibonacci r x y p q m (n + 1)
      = p * x * pqRFibonacci r (p * x) (q * y) p q m n
        + q * y * pqRFibonacci r (p * x) (q ^ (m + 1) * y) p q m (n - r)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 2, Equation (4).
Uses 1

TheoremSecond recurrence for the \((p,q)\)-analogue of the \(r\)-Fibonacci polynomial#●ready to prove

Let \(R\) be a commutative semiring, let \(r \ge 1\), \(n \ge r\) and \(m\) be natural numbers and let \(x, y, p, q \in R\). Write \(U_j(x, y)\) for \(U^{(r)}_j(x, y, p, q, m)\), the \((p,q)\)-analogue of the \(r\)-Fibonacci polynomial with \(r\), \(p\), \(q\) and \(m\) fixed. Then

\[U_{n+1}(x, y) = p x \, U_n(p x, p y) + q y \, U_{n-r}(q x, q^{m+1} y).\]
public theorem_wanted pqRFibonacci_succ_eq_right {R : Type*} [CommSemiring R] {r n : ℕ}
    (hr : 1 ≤ r) (hn : r ≤ n) (x y p q : R) (m : ℕ) :
    pqRFibonacci r x y p q m (n + 1)
      = p * x * pqRFibonacci r (p * x) (p * y) p q m n
        + q * y * pqRFibonacci r (q * x) (q ^ (m + 1) * y) p q m (n - r)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 2, Equation (5).
Uses 1

DefinitionFirst-kind (p, q)-Lucas polynomials#✓fully proved

Let \(R\) be a commutative semiring, let \(r\), \(s\) and \(m\) be natural numbers, let \(x, y, p, q \in R\), and write \([a, b]_{p,q}\) for the \((p, q)\)-binomial coefficients. The first-kind \((p, q)\)-analogue \(V^{(r,s)}_n(x, y, p, q, m)\) of the \(r\)-Lucas polynomial of type \(s\), written \(V_n\), is defined by \(V_0 = s + 1\) and, for \(n \ge 1\),

\[ V_n = \sum_{k=0}^{\lfloor n/(r+1) \rfloor} p^{\binom{n-(r+1)k}{2}} q^{(m+1)\binom{k}{2}}\, c_{n,k}\,x^{n-(r+1)k} y^k, \]

where \(c_{n,0} = [n, 0]_{p,q}\) and, for \(k \ge 1\), \(c_{n,k} = [n - rk, k]_{p,q} + s\,p^{n-(r+1)k} [n - rk - 1, k - 1]_{p,q}\).

The paper writes \(c_{n,k}\) as \([n - rk, k]_{p,q} (1 + s\,p^{n-(r+1)k} [k]_{p,q} / [n - rk]_{p,q})\), with \([a]_{p,q}\) the \((p, q)\)-numbers; the division-free form is the first step of the paper's proof of Theorem 3 and agrees with it whenever \([n - rk]_{p,q} \ne 0\).

public def pqRLucas {R : Type*} [CommSemiring R] (r s : ℕ) (x y p q : R) (m : ℕ) : ℕ → R
  | 0 => (s : R) + 1
  | n + 1 => ∑ k ∈ Finset.range ((n + 1) / (r + 1) + 1),
      p ^ (n + 1 - (r + 1) * k).choose 2 * q ^ ((m + 1) * k.choose 2) *
        (pqBinomial p q (n + 1 - r * k) k +
          if k = 0 then 0 else s * p ^ (n + 1 - (r + 1) * k) * pqBinomial p q (n - r * k) (k - 1)) *
        x ^ (n + 1 - (r + 1) * k) * y ^ k
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 2.
Uses 1 · Used by 4

TheoremInitial value of the first-kind (p, q)-Lucas polynomials#✓fully proved

Let \(R\) be a commutative semiring, let \(r\), \(s\) and \(m\) be natural numbers, and let \(x, y, p, q \in R\). The first-kind \((p, q)\)-analogue \(V^{(r,s)}_n(x, y, p, q, m)\) of the \(r\)-Lucas polynomial of type \(s\) satisfies \(V^{(r,s)}_0(x, y, p, q, m) = s + 1\).

public theorem pqRLucas_zero {R : Type*} [CommSemiring R] (r s : ℕ) (x y p q : R) (m : ℕ) :
    pqRLucas r s x y p q m 0 = (s : R) + 1
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 2.
Uses 1

DefinitionSecond-kind (p, q)-Lucas polynomials#✓fully proved

Let \(K\) be a field, let \(r\), \(s\) and \(m\) be natural numbers, let \(x, y, p, q \in K\), and write \([a, b]_{p,q}\) for the \((p, q)\)-binomial coefficients. The second-kind \((p, q)\)-analogue \(\mathbb{V}^{(r,s)}_n(x, y, p, q, m)\) of the \(r\)-Lucas polynomial of type \(s\), written \(\mathbb{V}_n\), is defined by \(\mathbb{V}_0 = s + 1\) and, for \(n \ge 1\),

\[ \mathbb{V}_n = \sum_{k=0}^{\lfloor n/(r+1) \rfloor} \frac{p^{\binom{n-(r+1)k}{2}} q^{\binom{k+1}{2} + m\binom{k}{2}}}{p^k}\, d_{n,k}\,x^{n-(r+1)k} y^k, \]

where \(d_{n,0} = [n, 0]_{p,q}\) and, for \(k \ge 1\), \(d_{n,k} = [n - rk, k]_{p,q} + s\,q^{n-(r+1)k} [n - rk - 1, k - 1]_{p,q}\). Division by \(0\) is read as \(0\), so for \(p = 0\) every term with \(k \ge 1\) vanishes.

The paper writes \(d_{n,k}\) as \([n - rk, k]_{p,q} (1 + s\,q^{n-(r+1)k} [k]_{p,q} / [n - rk]_{p,q})\), with \([a]_{p,q}\) the \((p, q)\)-numbers; the two agree whenever \([n - rk]_{p,q} \ne 0\).

public def pqRLucasSecond {K : Type*} [Field K] (r s : ℕ) (x y p q : K) (m : ℕ) : ℕ → K
  | 0 => (s : K) + 1
  | n + 1 => ∑ k ∈ Finset.range ((n + 1) / (r + 1) + 1),
      p ^ (n + 1 - (r + 1) * k).choose 2 * q ^ ((k + 1).choose 2 + m * k.choose 2) / p ^ k *
        (pqBinomial p q (n + 1 - r * k) k +
          if k = 0 then 0 else s * q ^ (n + 1 - (r + 1) * k) * pqBinomial p q (n - r * k) (k - 1)) *
        x ^ (n + 1 - (r + 1) * k) * y ^ k
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 2.
Uses 1 · Used by 4

TheoremInitial value of the second-kind (p, q)-Lucas polynomials#✓fully proved

Let \(K\) be a field, let \(r\), \(s\) and \(m\) be natural numbers, and let \(x, y, p, q \in K\). The second-kind \((p, q)\)-analogue \(\mathbb{V}^{(r,s)}_n(x, y, p, q, m)\) of the \(r\)-Lucas polynomial of type \(s\) satisfies \(\mathbb{V}^{(r,s)}_0(x, y, p, q, m) = s + 1\).

public theorem pqRLucasSecond_zero {K : Type*} [Field K] (r s : ℕ) (x y p q : K) (m : ℕ) :
    pqRLucasSecond r s x y p q m 0 = (s : K) + 1
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 2.
Uses 1

TheoremSplitting formula for first-kind (p, q)-Lucas polynomials#●ready to prove

Let \(K\) be a field, let \(r \ge 1\), \(s\) and \(m\) be natural numbers, and let \(x, y, p, q \in K\) with \(p \ne 0\) and \(q \ne 0\). Write \(U_n(x, y)\) for the \((p, q)\)-analogue of the \(r\)-Fibonacci polynomial with parameters \(p, q, m\), and \(V_n(x, y)\) for the first-kind \((p, q)\)-analogue of the \(r\)-Lucas polynomial of type \(s\) with the same parameters. Then for every natural number \(n \ge r\),

\[ V_n(x, y) = U_{n+1}(x/p, y/q) + s\,y\,U_{n-r}(x, y q^m). \]

The paper takes positive \(s\); the identity holds for every natural number \(s\).

public theorem_wanted pqRLucas_eq_pqRFibonacci_add {K : Type*} [Field K] {r n : ℕ}
    (hr : 1 ≤ r) (hn : r ≤ n) (s : ℕ) (x y p q : K) (hp : p ≠ 0) (hq : q ≠ 0) (m : ℕ) :
    pqRLucas r s x y p q m n
      = pqRFibonacci r (x / p) (y / q) p q m (n + 1)
        + s * y * pqRFibonacci r x (y * q ^ m) p q m (n - r)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 3, part 1.
Uses 2

TheoremSplitting formula for second-kind (p, q)-Lucas polynomials#●ready to prove

Let \(K\) be a field, let \(r \ge 1\), \(s\) and \(m\) be natural numbers, and let \(x, y, p, q \in K\) with \(p \ne 0\) and \(q \ne 0\). Write \(U_n(x, y)\) for the \((p, q)\)-analogue of the \(r\)-Fibonacci polynomial with parameters \(p, q, m\), and \(\mathbb{V}_n(x, y)\) for the second-kind \((p, q)\)-analogue of the \(r\)-Lucas polynomial of type \(s\) with the same parameters. Then for every natural number \(n \ge r\),

\[ \mathbb{V}_n(x, y) = U_{n+1}(x/p, y/p) + s\,y\,(q/p)^{n-r}\,U_{n-r}(x, y p^r q^m / q^r). \]
public theorem_wanted pqRLucasSecond_eq_pqRFibonacci_add {K : Type*} [Field K] {r n : ℕ}
    (hr : 1 ≤ r) (hn : r ≤ n) (s : ℕ) (x y p q : K) (hp : p ≠ 0) (hq : q ≠ 0) (m : ℕ) :
    pqRLucasSecond r s x y p q m n
      = pqRFibonacci r (x / p) (y / p) p q m (n + 1)
        + s * y * (q / p) ^ (n - r) * pqRFibonacci r x (y * p ^ r * q ^ m / q ^ r) p q m (n - r)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 3, part 1.
Uses 2

TheoremFirst-kind (p, q)-Lucas polynomials as a difference of (p, q)-Fibonacci polynomials#●ready to prove

Let \(K\) be a field, let \(r \ge 1\), \(s\) and \(m\) be natural numbers, and let \(x, y, p, q \in K\) with \(p \ne 0\) and \(q \ne 0\). Write \(U_n(x, y)\) for the \((p, q)\)-analogue of the \(r\)-Fibonacci polynomial with parameters \(p, q, m\), and \(V_n(x, y)\) for the first-kind \((p, q)\)-analogue of the \(r\)-Lucas polynomial of type \(s\) with the same parameters. Then for every natural number \(n\),

\[ V_n(x, y) = (s + 1)\,U_{n+1}(x/p, y/q) - s\,x\,U_n(x, y). \]
public theorem_wanted pqRLucas_eq_succ_mul_sub {K : Type*} [Field K] {r : ℕ} (hr : 1 ≤ r)
    (s : ℕ) (x y p q : K) (hp : p ≠ 0) (hq : q ≠ 0) (m n : ℕ) :
    pqRLucas r s x y p q m n
      = (s + 1) * pqRFibonacci r (x / p) (y / q) p q m (n + 1)
        - s * x * pqRFibonacci r x y p q m n
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 3, part 2.
Uses 2

TheoremSecond-kind (p, q)-Lucas polynomials as a difference of (p, q)-Fibonacci polynomials#●ready to prove

Let \(K\) be a field, let \(r \ge 1\), \(s\) and \(m\) be natural numbers, and let \(x, y, p, q \in K\) with \(p \ne 0\). Write \(U_n(x, y)\) for the \((p, q)\)-analogue of the \(r\)-Fibonacci polynomial with parameters \(p, q, m\), and \(\mathbb{V}_n(x, y)\) for the second-kind \((p, q)\)-analogue of the \(r\)-Lucas polynomial of type \(s\) with the same parameters. Then for every natural number \(n\),

\[ \mathbb{V}_n(x, y) = (s + 1)\,U_{n+1}(x/p, y/p) - s\,x\,U_n(x, y). \]
public theorem_wanted pqRLucasSecond_eq_succ_mul_sub {K : Type*} [Field K] {r : ℕ}
    (hr : 1 ≤ r) (s : ℕ) (x y p q : K) (hp : p ≠ 0) (m n : ℕ) :
    pqRLucasSecond r s x y p q m n
      = (s + 1) * pqRFibonacci r (x / p) (y / p) p q m (n + 1)
        - s * x * pqRFibonacci r x y p q m n
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 3, part 2.
Uses 2

TheoremFirst-kind \((p,q)\)-analogue of the \(r\)-Lucas polynomial via two \(r\)-Fibonacci terms#●ready to prove

Let \(R\) be a commutative semiring, let \(r \ge 1\), \(n \ge r\), \(s\) and \(m\) be natural numbers and let \(x, y, p, q \in R\). Write \(U_j(x, y)\) for the \((p,q)\)-analogue \(U^{(r)}_j(x, y, p, q, m)\) of the \(r\)-Fibonacci polynomial and \(V_n(x, y)\) for the first-kind \((p,q)\)-analogue \(V^{(r,s)}_n(x, y, p, q, m)\) of the \(r\)-Lucas polynomial of type \(s\), with \(r\), \(s\), \(p\), \(q\) and \(m\) fixed. Then

\[V_n(x, y) = x \, U_n(x, y) + (1 + s) \, y \, U_{n-r}(x, y q^m).\]
public theorem_wanted pqRLucas_eq_mul_add {R : Type*} [CommSemiring R] {r n : ℕ}
    (hr : 1 ≤ r) (hn : r ≤ n) (s : ℕ) (x y p q : R) (m : ℕ) :
    pqRLucas r s x y p q m n
      = x * pqRFibonacci r x y p q m n
        + (1 + s) * y * pqRFibonacci r x (y * q ^ m) p q m (n - r)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 3, part 3.
Uses 2

TheoremRecurrence-type formula for second-kind (p, q)-Lucas polynomials#●ready to prove

Let \(K\) be a field, let \(r \ge 1\), \(s\) and \(m\) be natural numbers, and let \(x, y, p, q \in K\) with \(p \ne 0\) and \(q \ne 0\). Write \(U_n(x, y)\) for the \((p, q)\)-analogue of the \(r\)-Fibonacci polynomial with parameters \(p, q, m\), and \(\mathbb{V}_n(x, y)\) for the second-kind \((p, q)\)-analogue of the \(r\)-Lucas polynomial of type \(s\) with the same parameters. Then for every natural number \(n \ge r\),

\[ \mathbb{V}_n(x, y) = x\,U_n(x, y) + (1 + s)\,(q/p)^{n-r}\,y\,U_{n-r}(x, y p^r q^m / q^r). \]

The paper writes the second argument as \(y p^r q^{m-r}\) and drops \(p, q\) from the argument list \((x, y p^r q^{m-r}, m)\); they are unchanged.

public theorem_wanted pqRLucasSecond_eq_mul_add {K : Type*} [Field K] {r n : ℕ}
    (hr : 1 ≤ r) (hn : r ≤ n) (s : ℕ) (x y p q : K) (hp : p ≠ 0) (hq : q ≠ 0) (m : ℕ) :
    pqRLucasSecond r s x y p q m n
      = x * pqRFibonacci r x y p q m n
        + (1 + s) * (q / p) ^ (n - r) * y *
            pqRFibonacci r x (y * p ^ r * q ^ m / q ^ r) p q m (n - r)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 3, part 3.
Uses 2

TheoremGeometrically weighted sum of \(r\)-Lucas polynomials#●ready to prove

Let \(K\) be a field, let \(r \ge 1\) and \(s\) be natural numbers, let \(x, y, m \in K\) with \(1 - m x - m^{r+1} y \ne 0\), and write \(V_j\) for the \(r\)-Lucas polynomial of type \(s\), \(V_j = V^{(r,s)}_j(x, y)\). Then for every natural number \(n \ge r - 1\),

\[ \sum_{j=1}^{n} m^j V_j = \frac{1 + s m^{r+1} y - m^{n+1} V_{n+1} - m^{r+1} y \sum_{j=1}^{r} m^{n-r+j} V_{n-r+j}} {1 - m x - m^{r+1} y} - 1. \]

The paper takes \(m\) a positive integer (here any element of \(K\)) and states no range for \(n\); for \(n < r - 1\) the right side involves \(V\) at negative indices.

public theorem_wanted sum_pow_mul_rLucasPolynomial {K : Type*} [Field K] {r n : ℕ}
    (hr : 1 ≤ r) (hn : r ≤ n + 1) (s : ℕ) (x y m : K) (hP : 1 - m * x - m ^ (r + 1) * y ≠ 0) :
    ∑ j ∈ Finset.Icc 1 n, m ^ j * rLucasPolynomial r s x y j
      = (1 + s * m ^ (r + 1) * y - m ^ (n + 1) * rLucasPolynomial r s x y (n + 1)
          - m ^ (r + 1) * y * ∑ j ∈ Finset.Icc 1 r,
              m ^ (n + j - r) * rLucasPolynomial r s x y (n + j - r))
        / (1 - m * x - m ^ (r + 1) * y) - 1
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 5.
Uses 1

TheoremSum of the first \(n\) \(r\)-Lucas polynomials#●ready to prove

Let \(K\) be a field, let \(r \ge 1\) and \(s\) be natural numbers, let \(x, y \in K\) with \(1 - x - y \ne 0\), and write \(V_j\) for the \(r\)-Lucas polynomial of type \(s\), \(V_j = V^{(r,s)}_j(x, y)\). Then for every natural number \(n\) with \(n \ge 1\) and \(n \ge r - 1\),

\[ \sum_{j=1}^{n} V_j = \frac{1 + s y - V_{n+1} - y \sum_{j=1}^{r} V_{n-r+j}}{1 - x - y} - 1. \]

The paper states \(n \ge 1\); for \(n < r - 1\) the right side involves \(V\) at negative indices.

public theorem_wanted sum_rLucasPolynomial {K : Type*} [Field K] {r n : ℕ} (hr : 1 ≤ r)
    (hn1 : 1 ≤ n) (hn : r ≤ n + 1) (s : ℕ) (x y : K) (hP : 1 - x - y ≠ 0) :
    ∑ j ∈ Finset.Icc 1 n, rLucasPolynomial r s x y j
      = (1 + s * y - rLucasPolynomial r s x y (n + 1)
          - y * ∑ j ∈ Finset.Icc 1 r, rLucasPolynomial r s x y (n + j - r))
        / (1 - x - y) - 1
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Corollary 4.
Uses 1

TheoremAlternating sum of \(r\)-Lucas polynomials#●ready to prove

Let \(K\) be a field, let \(r \ge 1\) and \(s\) be natural numbers, let \(x, y \in K\) with \(1 + x + (-1)^r y \ne 0\), and write \(V_j\) for the \(r\)-Lucas polynomial of type \(s\), \(V_j = V^{(r,s)}_j(x, y)\). Then for every natural number \(n \ge r - 1\),

\[ \sum_{j=1}^{n} (-1)^j V_j = \frac{1 - (-1)^r s y + (-1)^n V_{n+1} + (-1)^r y \sum_{j=1}^{r} (-1)^{n-r+j} V_{n-r+j}} {1 + x + (-1)^r y} - 1. \]

For \(n < r - 1\) the right side involves \(V\) at negative indices.

public theorem_wanted sum_neg_one_pow_mul_rLucasPolynomial {K : Type*} [Field K] {r n : ℕ}
    (hr : 1 ≤ r) (hn : r ≤ n + 1) (s : ℕ) (x y : K) (hP : 1 + x + (-1) ^ r * y ≠ 0) :
    ∑ j ∈ Finset.Icc 1 n, (-1) ^ j * rLucasPolynomial r s x y j
      = (1 - (-1) ^ r * s * y + (-1) ^ n * rLucasPolynomial r s x y (n + 1)
          + (-1) ^ r * y * ∑ j ∈ Finset.Icc 1 r,
              (-1) ^ (n + j - r) * rLucasPolynomial r s x y (n + j - r))
        / (1 + x + (-1) ^ r * y) - 1
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Corollary 5.
Uses 1

DefinitionThe (-r)-Fibonacci polynomials#✓fully proved

Let \(A\) be a commutative ring, \(r\) a natural number, \(x \in A\) and \(y\) a unit of \(A\). The values \(U_n = U^{(-r)}_n(x, y) \in A\) of the \((-r)\)-Fibonacci polynomials are defined for \(n \ge 0\) by \(U_0 = 0\), \(U_1 = 1\), \(U_n = 0\) for \(2 \le n \le r\), and, when \(r \ge 1\), \(U_{n+1} = y^{-1} U_{n-r+1} - y^{-1} x U_{n-r}\) for \(n \ge r\). When \(r = 0\), \(U_n = 0\) for every \(n \ge 2\).

The paper takes \(r \ge 2\); the case \(r = 0\) only makes the definition total.

public def negRFibonacci {A : Type*} [CommRing A] (r : ℕ) (x : A) (y : Aˣ) : ℕ → A
  | 0 => 0
  | n + 1 =>
    if n = 0 then 1
    else if n < r ∨ r = 0 then 0
    else (↑y⁻¹ : A) * negRFibonacci r x y (n - r + 1) - ↑y⁻¹ * x * negRFibonacci r x y (n - r)
termination_by n => n
decreasing_by all_goals omega
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 3.
Used by 9

TheoremThe zeroth (-r)-Fibonacci polynomial#✓fully proved

For every commutative ring \(A\), natural number \(r\), \(x \in A\) and unit \(y\) of \(A\), the \((-r)\)-Fibonacci polynomial satisfies \(U^{(-r)}_0(x, y) = 0\).

public theorem negRFibonacci_zero {A : Type*} [CommRing A] (r : ℕ) (x : A) (y : Aˣ) :
    negRFibonacci r x y 0 = 0
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 3.
Uses 1

TheoremThe first (-r)-Fibonacci polynomial#✓fully proved

For every commutative ring \(A\), natural number \(r\), \(x \in A\) and unit \(y\) of \(A\), the \((-r)\)-Fibonacci polynomial satisfies \(U^{(-r)}_1(x, y) = 1\).

public theorem negRFibonacci_one {A : Type*} [CommRing A] (r : ℕ) (x : A) (y : Aˣ) :
    negRFibonacci r x y 1 = 1
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 3.
Uses 1

TheoremThe (-r)-Fibonacci polynomials vanish at indices 2 to r#✓fully proved

Let \(A\) be a commutative ring, \(x \in A\) and \(y\) a unit of \(A\). For natural numbers \(r\) and \(n\) with \(1 \le n < r\), the \((-r)\)-Fibonacci polynomial satisfies \(U^{(-r)}_{n+1}(x, y) = 0\).

public theorem negRFibonacci_succ_of_lt {A : Type*} [CommRing A] {r n : ℕ} (x : A) (y : Aˣ)
    (h1 : 1 ≤ n) (h : n < r) : negRFibonacci r x y (n + 1) = 0
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 3.
Uses 1

TheoremRecurrence for the (-r)-Fibonacci polynomials#✓fully proved

Let \(A\) be a commutative ring, \(x \in A\) and \(y\) a unit of \(A\), let \(r\) and \(n\) be natural numbers with \(1 \le r \le n\), and write \(U_m = U^{(-r)}_m(x, y)\). Then \(U_{n+1} = y^{-1} U_{n-r+1} - y^{-1} x U_{n-r}\).

public theorem negRFibonacci_succ_of_le {A : Type*} [CommRing A] {r n : ℕ} (x : A) (y : Aˣ)
    (hr : 1 ≤ r) (h : r ≤ n) :
    negRFibonacci r x y (n + 1)
      = ↑y⁻¹ * negRFibonacci r x y (n - r + 1) - ↑y⁻¹ * x * negRFibonacci r x y (n - r)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 3, Equation (7).
Uses 1

DefinitionCompanion sequence of the \((-r)\)-Fibonacci polynomials#✓fully proved

Let \(A\) be a commutative ring, \(r\) a natural number, \(x \in A\) and \(y\) a unit of \(A\). The companion sequence \(V^{(-r)}_n(x, y) \in A\), \(n \ge 0\), is defined by \(V^{(-r)}_0 = r + 1\) and, if \(r \ge 1\), by \(V^{(-r)}_n = 0\) for \(1 \le n < r\), \(V^{(-r)}_r = r y^{-1}\) and \(V^{(-r)}_{n+1} = y^{-1} V^{(-r)}_{n-r+1} - y^{-1} x V^{(-r)}_{n-r}\) for \(n \ge r\). If \(r = 0\), then \(V^{(-r)}_n = 0\) for all \(n \ge 1\).

The paper takes \(r \ge 2\); the case \(r = 0\) only makes the definition total.

public def negRLucas {A : Type*} [CommRing A] (r : ℕ) (x : A) (y : Aˣ) : ℕ → A
  | 0 => (r : A) + 1
  | n + 1 =>
    if r = 0 ∨ n + 1 < r then 0
    else if n + 1 = r then (r : A) * ↑y⁻¹
    else (↑y⁻¹ : A) * negRLucas r x y (n - r + 1) - ↑y⁻¹ * x * negRLucas r x y (n - r)
termination_by n => n
decreasing_by all_goals omega
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 4.
Used by 9

TheoremZeroth term of the companion sequence \(V^{(-r)}\)#✓fully proved

Let \(A\) be a commutative ring, \(r\) a natural number, \(x \in A\) and \(y\) a unit of \(A\). The companion sequence of the \((-r)\)-Fibonacci polynomials satisfies \(V^{(-r)}_0(x, y) = r + 1\).

public theorem negRLucas_zero {A : Type*} [CommRing A] (r : ℕ) (x : A) (y : Aˣ) :
    negRLucas r x y 0 = (r : A) + 1
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 4.
Uses 1

TheoremVanishing initial terms of the companion sequence \(V^{(-r)}\)#✓fully proved

Let \(A\) be a commutative ring, \(x \in A\) and \(y\) a unit of \(A\). For natural numbers \(r\) and \(n\) with \(n + 1 < r\), the companion sequence of the \((-r)\)-Fibonacci polynomials satisfies \(V^{(-r)}_{n+1}(x, y) = 0\); that is, \(V^{(-r)}_m(x, y) = 0\) for \(1 \le m < r\).

public theorem negRLucas_succ_of_lt {A : Type*} [CommRing A] {r n : ℕ} (x : A) (y : Aˣ)
    (h : n + 1 < r) : negRLucas r x y (n + 1) = 0
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 4.
Uses 1

TheoremThe \(r\)-th term of the companion sequence \(V^{(-r)}\)#✓fully proved

Let \(A\) be a commutative ring, \(r \ge 1\) a natural number, \(x \in A\) and \(y\) a unit of \(A\). The companion sequence of the \((-r)\)-Fibonacci polynomials satisfies \(V^{(-r)}_r(x, y) = r y^{-1}\).

public theorem negRLucas_self {A : Type*} [CommRing A] {r : ℕ} (x : A) (y : Aˣ)
    (hr : 1 ≤ r) : negRLucas r x y r = (r : A) * ↑y⁻¹
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 4.
Uses 1

TheoremRecurrence for the companion sequence \(V^{(-r)}\)#✓fully proved

Let \(A\) be a commutative ring, \(x \in A\) and \(y\) a unit of \(A\). For natural numbers \(r\) and \(n\) with \(1 \le r \le n\), the companion sequence of the \((-r)\)-Fibonacci polynomials satisfies \(V^{(-r)}_{n+1}(x, y) = y^{-1} V^{(-r)}_{n-r+1}(x, y) - y^{-1} x V^{(-r)}_{n-r}(x, y)\).

public theorem negRLucas_succ_of_le {A : Type*} [CommRing A] {r n : ℕ} (x : A) (y : Aˣ)
    (hr : 1 ≤ r) (h : r ≤ n) :
    negRLucas r x y (n + 1)
      = ↑y⁻¹ * negRLucas r x y (n - r + 1) - ↑y⁻¹ * x * negRLucas r x y (n - r)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Definition 4.
Uses 1

TheoremExplicit sum for the (-r)-Fibonacci polynomials#●ready to prove

Let \(A\) be a commutative ring, \(r \ge 2\), \(x \in A\) and \(y\) a unit of \(A\), and write \(U_n = U^{(-r)}_n(x, y)\). For every \(n \ge 0\),

\[U_{n+1} = \sum_{k=0}^{\lfloor n/r \rfloor} \binom{k}{n - rk} y^{-k} (-x)^{n-rk},\]

where \(\binom{a}{b} = 0\) for \(b > a\), so the terms with \((r+1) k < n\) vanish.

The paper states \(n \ge 1\) and \(x\) a unit; the identity holds for \(n = 0\) and every \(x\).

public theorem_wanted negRFibonacci_succ_eq_sum {A : Type*} [CommRing A] {r : ℕ}
    (hr : 2 ≤ r) (x : A) (y : Aˣ) (n : ℕ) :
    negRFibonacci r x y (n + 1)
      = ∑ k ∈ Finset.range (n / r + 1),
          ((k.choose (n - r * k) : ℕ) : A) * (↑y⁻¹ : A) ^ k * (-x) ^ (n - r * k)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 6.
Uses 1

TheoremExplicit sum for the (-r)-Fibonacci polynomials, second form#●ready to prove

Let \(A\) be a commutative ring, \(r \ge 2\), \(x \in A\) and \(y\) a unit of \(A\), and write \(U_n = U^{(-r)}_n(x, y)\). For every \(n \ge 0\), \(U_{n+1}\) is the sum of \(\binom{(n-k)/r}{k} y^{-(n-k)/r} (-x)^k\) over the integers \(k\) with \(0 \le k \le \lfloor n/r \rfloor\) and \(r \mid n - k\).

public theorem_wanted negRFibonacci_succ_eq_sum' {A : Type*} [CommRing A] {r : ℕ}
    (hr : 2 ≤ r) (x : A) (y : Aˣ) (n : ℕ) :
    negRFibonacci r x y (n + 1)
      = ∑ k ∈ (Finset.range (n / r + 1)).filter (fun k => r ∣ n - k),
          ((((n - k) / r).choose k : ℕ) : A) * (↑y⁻¹ : A) ^ ((n - k) / r) * (-x) ^ k
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 6.
Uses 1

TheoremThe companion sequence in terms of the \((-r)\)-Fibonacci polynomials#●ready to prove

Let \(A\) be a commutative ring, \(r \ge 2\) a natural number, \(x \in A\) and \(y\) a unit of \(A\), and let \(U^{(-r)}_n(x, y)\) be the \((-r)\)-Fibonacci polynomials and \(V^{(-r)}_n(x, y)\) their companion sequence. For every natural number \(n \ge r\), \(V^{(-r)}_n(x, y) = r \, U^{(-r)}_{n+1}(x, y) - x y^{-1} U^{(-r)}_{n-r}(x, y)\).

The paper takes \(x\) a unit; the identity holds for every \(x\).

public theorem_wanted negRLucas_eq {A : Type*} [CommRing A] {r n : ℕ} (hr : 2 ≤ r)
    (hn : r ≤ n) (x : A) (y : Aˣ) :
    negRLucas r x y n = r * negRFibonacci r x y (n + 1) - x * ↑y⁻¹ * negRFibonacci r x y (n - r)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Proposition 1, Equation (8).
Uses 2

TheoremExplicit sum for the companion sequence \(V^{(-r)}\)#●ready to prove

Let \(K\) be a field of characteristic zero, \(r \ge 2\) and \(n \ge 1\) natural numbers, \(x \in K\) and \(y\) a unit of \(K\), and let \(V^{(-r)}_n(x, y)\) be the companion sequence of the \((-r)\)-Fibonacci polynomials. Then

\[V^{(-r)}_n(x, y) = \sum_{k} \frac{n}{n - rk} \binom{k - 1}{n - 1 - rk} y^{-k} \, (-x)^{n - rk} + \varepsilon,\]

where the sum runs over the natural numbers \(k \ge 1\) with \(rk < n\), \(\binom{a}{b}\) is the binomial coefficient of natural numbers (zero when \(b > a\)), and \(\varepsilon = r y^{-n/r}\) if \(r\) divides \(n\) and \(\varepsilon = 0\) otherwise.

The paper's range \(\lfloor n/(r+1) \rfloor \le k \le \lfloor n/r \rfloor\) also admits \(k = n/r\) (division by zero; that term is the \(\varepsilon\) term) and, for \(n \le r\), \(k = 0\), whose term \(\binom{-1}{n-1}\) must be read as \(0\) (with \(\binom{-1}{j} = (-1)^j\), \(r = 2\), \(n = 1\) gives \(-x\) instead of \(V^{(-r)}_1 = 0\)); both are excluded.

public theorem_wanted negRLucas_eq_sum {K : Type*} [Field K] [CharZero K] {r n : ℕ}
    (hr : 2 ≤ r) (hn : 1 ≤ n) (x : K) (y : Kˣ) :
    negRLucas r x y n
      = ∑ k ∈ (Finset.Icc 1 (n / r)).filter (fun k => r * k < n),
          (n : K) / ((n : K) - ((r * k : ℕ) : K)) * ((k - 1).choose (n - 1 - r * k) : K) *
            (↑y⁻¹ : K) ^ k * (-x) ^ (n - r * k)
        + if r ∣ n then (r : K) * (↑y⁻¹ : K) ^ (n / r) else 0
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 7.
Uses 1

TheoremExplicit sum for the (-r)-Lucas polynomials, second form#●ready to prove

Let \(K\) be a field of characteristic zero, \(r \ge 2\), \(x \in K\) and \(y \in K\) nonzero, and write \(V_n = V^{(-r)}_n(x, y)\) for the companion \((-r)\)-Lucas sequence. For every \(n \ge 1\), \(V_n\) is the sum of

\[\frac{n}{k+1} \binom{(n-1-r-k)/r}{k} y^{-(n-1-k)/r} (-x)^{k+1}\]

over the integers \(k\) with \(0 \le k \le \lfloor n/(r+1) \rfloor - 1\) and \(r \mid n - k - 1\), plus \(r y^{-n/r}\) when \(r \mid n\).

public theorem_wanted negRLucas_eq_sum' {K : Type*} [Field K] [CharZero K] {r n : ℕ}
    (hr : 2 ≤ r) (hn : 1 ≤ n) (x : K) (y : Kˣ) :
    negRLucas r x y n
      = ∑ k ∈ (Finset.range (n / (r + 1))).filter (fun k => r ∣ n - k - 1),
          (n : K) / (k + 1) * (((n - 1 - r - k) / r).choose k : K) *
            (↑y⁻¹ : K) ^ ((n - 1 - k) / r) * (-x) ^ (k + 1)
        + if r ∣ n then (r : K) * (↑y⁻¹ : K) ^ (n / r) else 0
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 7.
Uses 1

TheoremGenerating function of the (-r)-Fibonacci polynomials#●ready to prove

Let \(A\) be a commutative ring, \(r \ge 2\), \(x \in A\) and \(y\) a unit of \(A\), and write \(U_n = U^{(-r)}_n(x, y)\). In the ring \(A[[z]]\) of formal power series,

\[\left(1 - y^{-1} z^r + x y^{-1} z^{r+1}\right) \sum_{n \ge 0} U_{n+1} z^n = 1.\]

The paper states the identity for complex \(z\); convergence is not part of this statement.

public theorem_wanted negRFibonacci_generatingFunction {A : Type*} [CommRing A] {r : ℕ}
    (hr : 2 ≤ r) (x : A) (y : Aˣ) :
    (1 - PowerSeries.C (↑y⁻¹ : A) * PowerSeries.X ^ r
        + PowerSeries.C (x * ↑y⁻¹) * PowerSeries.X ^ (r + 1)) *
      PowerSeries.mk (fun n => negRFibonacci r x y (n + 1)) = 1
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 8.
Uses 1

TheoremGenerating function of the companion sequence \(V^{(-r)}\)#●ready to prove

Let \(A\) be a commutative ring, \(r \ge 2\) a natural number, \(x \in A\) and \(y\) a unit of \(A\), and let \(V^{(-r)}_n(x, y)\) be the companion sequence of the \((-r)\)-Fibonacci polynomials. In the ring \(A[[z]]\) of formal power series,

\[(1 - y^{-1} z^r + x y^{-1} z^{r+1}) \sum_{n \ge 0} V^{(-r)}_n(x, y) z^n = r + 1 - y^{-1} z^r.\]
public theorem_wanted negRLucas_generatingFunction {A : Type*} [CommRing A] {r : ℕ}
    (hr : 2 ≤ r) (x : A) (y : Aˣ) :
    (1 - PowerSeries.C (↑y⁻¹ : A) * PowerSeries.X ^ r
        + PowerSeries.C (x * ↑y⁻¹) * PowerSeries.X ^ (r + 1)) *
      PowerSeries.mk (fun n => negRLucas r x y n)
      = PowerSeries.C ((r : A) + 1) - PowerSeries.C (↑y⁻¹ : A) * PowerSeries.X ^ r
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 8.
Uses 1

TheoremSeparability of \(t^{r+1} - y^{-1} t + x y^{-1}\)#●ready to prove

Let \(K\) be a field of characteristic zero, let \(r \ge 1\), let \(x \in K\) and let \(y \in K\) be nonzero with \((r+1)^{r+1} x^r y \ne r^r\). Then the polynomial \(t^{r+1} - y^{-1} t + x y^{-1} \in K[t]\) is separable, that is, coprime to its derivative; equivalently, it has no multiple zeros in an algebraic closure of \(K\).

This corrects Lemma 1 of the paper. The printed hypothesis \(x \ne \frac{r}{r+1} ((r+1) y)^{-1/r}\) names only one \(r\)-th root and is insufficient: \(r = 2\), \(y = 1/3\), \(x = -2/3\) satisfies it, since \(\frac{2}{3} \cdot 1^{-1/2} = \frac{2}{3}\), yet \(t^3 - 3t - 2 = (t + 1)^2 (t - 2)\). The paper's proof shows that a multiple zero \(\beta\) forces \(x = r\beta/(r+1)\) with \(\beta^r = 1/((r+1) y)\), that is, \((r+1)^{r+1} x^r y = r^r\), which is the branch-free condition used here.

public theorem_wanted separable_negRFibonacci_charPoly {K : Type*} [Field K] [CharZero K]
    {r : ℕ} (hr : 1 ≤ r) (x : K) (y : Kˣ)
    (h : ((r : K) + 1) ^ (r + 1) * x ^ r * y ≠ (r : K) ^ r) :
    (Polynomial.X ^ (r + 1) - Polynomial.C (↑y⁻¹ : K) * Polynomial.X
      + Polynomial.C (x * ↑y⁻¹) : Polynomial K).Separable
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Lemma 1.

TheoremBinet formula for the (-r)-Fibonacci polynomials#●ready to prove

Let \(K\) be a field, \(r \ge 2\), \(x \in K\) and \(y \in K\) nonzero, and write \(U_n = U^{(-r)}_n(x, y)\). Suppose that \(\beta_0, \dots, \beta_r \in K\) are pairwise distinct and that \(t^{r+1} - y^{-1} t + x y^{-1} = \prod_{k=0}^{r} (t - \beta_k)\) as polynomials over \(K\). Then for every \(n \ge 0\),

\[U_{n+1} = \sum_{k=0}^{r} \frac{\beta_k^{n+r+1}}{r \beta_k^{r+1} - x y^{-1}},\]

where a term whose denominator vanishes is read as \(0\).

The paper's hypothesis \(x \ne \frac{r}{r+1} ((r+1) y)^{-1/r}\) is replaced by the distinctness of the zeros (compare Lemma 1).

public theorem_wanted negRFibonacci_eq_sum_roots {K : Type*} [Field K] {r : ℕ} (hr : 2 ≤ r)
    (x : K) (y : Kˣ) (β : Fin (r + 1) → K) (hβ : Function.Injective β)
    (hP : (Polynomial.X ^ (r + 1) - Polynomial.C (↑y⁻¹ : K) * Polynomial.X
      + Polynomial.C (x * ↑y⁻¹) : Polynomial K) = ∏ k, (Polynomial.X - Polynomial.C (β k)))
    (n : ℕ) :
    negRFibonacci r x y (n + 1) = ∑ k, β k ^ (n + r + 1) / (r * β k ^ (r + 1) - x * ↑y⁻¹)
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 9.
Uses 1

TheoremBinet formula for the companion sequence \(V^{(-r)}\)#●ready to prove

Let \(K\) be a field, \(r \ge 2\) a natural number, \(x \in K\) and \(y\) a unit of \(K\), and let \(V^{(-r)}_n(x, y)\) be the companion sequence of the \((-r)\)-Fibonacci polynomials. Suppose \(\beta_0, \dots, \beta_r \in K\) are pairwise distinct and \(t^{r+1} - y^{-1} t + x y^{-1} = \prod_{k=0}^{r} (t - \beta_k)\) in the polynomial ring \(K[t]\). Then \(V^{(-r)}_n(x, y) = \sum_{k=0}^{r} \beta_k^n\) for every natural number \(n\).

public theorem_wanted negRLucas_eq_sum_roots {K : Type*} [Field K] {r : ℕ} (hr : 2 ≤ r)
    (x : K) (y : Kˣ) (β : Fin (r + 1) → K) (hβ : Function.Injective β)
    (hP : (Polynomial.X ^ (r + 1) - Polynomial.C (↑y⁻¹ : K) * Polynomial.X
      + Polynomial.C (x * ↑y⁻¹) : Polynomial K) = ∏ k, (Polynomial.X - Polynomial.C (β k)))
    (n : ℕ) :
    negRLucas r x y n = ∑ k, β k ^ n
Sources
  • Abbad and Belbachir, The r-Fibonacci Polynomial and Its Companion Sequences Linked with Some Classical Sequences, INTEGERS 25 (2025), Article A38, Theorem 9.
Uses 1
31 fully proved28 ready to prove

The dependency explorer needs JavaScript and its graph data.

Download the complete graph data.