Binomial sums with second-order recurrences

TheoremAlternating squared-binomial sum with Lucas sequences#●ready to prove

Let \(p\) be an integer and let \(U_n\) and \(V_n\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\) and both satisfy \(X_{j+2} = p X_{j+1} + X_j\). Then for every natural number \(n\),

\[\sum_{k=0}^{2n} (-1)^k \binom{2n}{k}^2 U_{2n-k} V_k = (-1)^n \binom{2n}{n} U_{2n}.\]
public theorem_wanted lucasU_lucasV_sum_choose_sq_alternating (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (2 * n + 1),
        ((2 * n).choose k : ℤ) ^ 2 * (-1) ^ k *
          lucasU ⟨p, -1, by decide⟩ (2 * n - k) * lucasV ⟨p, -1, by decide⟩ k =
      (-1) ^ n * ((2 * n).choose n : ℤ) * lucasU ⟨p, -1, by decide⟩ (2 * n)
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Theorem 2.1.
Uses 3

TheoremDixon-type alternating sum with Lucas sequences#●ready to prove

Let \(p\) be an integer and let \(U_n\) and \(V_n\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\) and both satisfy \(X_{j+2} = p X_{j+1} + X_j\). Then for every natural number \(n\),

\[\sum_{k=0}^{2n} (-1)^k \binom{2n}{k}^3 U_{2n-k} V_k = (-1)^n \binom{3n}{n} \binom{2n}{n} U_{2n},\]

where \(\binom{3n}{n} \binom{2n}{n} = (3n)!/(n!)^3\) is the trinomial coefficient.

public theorem_wanted lucasU_lucasV_sum_choose_cube_alternating (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (2 * n + 1),
        ((2 * n).choose k : ℤ) ^ 3 * (-1) ^ k *
          lucasU ⟨p, -1, by decide⟩ (2 * n - k) * lucasV ⟨p, -1, by decide⟩ k =
      (-1) ^ n * ((3 * n).choose n * (2 * n).choose n : ℕ) * lucasU ⟨p, -1, by decide⟩ (2 * n)
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Theorem 2.2.
Uses 3

TheoremSquared-binomial convolution of the Lucas sequences U and V#●ready to prove

Let \(p\) be an integer and let \(U\) and \(V\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so that \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\), and both satisfy \(X_{j+2} = p X_{j+1} + X_j\). Then for every natural number \(n\)

\[\sum_{k=0}^{n} \binom{n}{k}^2 U_{n-k} V_k = \binom{2n}{n} U_n.\]
public theorem_wanted lucasU_lucasV_sum_choose_sq (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (n + 1),
        (n.choose k : ℤ) ^ 2 * lucasU ⟨p, -1, by decide⟩ (n - k) * lucasV ⟨p, -1, by decide⟩ k =
      ((2 * n).choose n : ℤ) * lucasU ⟨p, -1, by decide⟩ n
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 2.3.
Uses 3

TheoremEven-index Catalan convolution with Lucas sequences#●ready to prove

Let \(p\) be an integer, let \(U_n\) and \(V_n\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\) and both satisfy \(X_{j+2} = p X_{j+1} + X_j\), and let \(C_j\) be the \(j\)-th Catalan number. Then for every natural number \(n\), in the rationals,

\[\sum_{k=0}^{n} C_{2n-2k} C_{2k} U_k V_{n-k} = \frac{2^{2n}}{2} \cdot \frac{n+2}{2n+1} C_{n+1} U_n.\]
public theorem_wanted lucasU_lucasV_sum_catalan_even (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (n + 1),
        (catalan (2 * n - 2 * k) * catalan (2 * k) : ℚ) *
          lucasU ⟨p, -1, by decide⟩ k * lucasV ⟨p, -1, by decide⟩ (n - k) =
      (2 : ℚ) ^ (2 * n) / 2 * ((n + 2 : ℚ) / (2 * n + 1)) * catalan (n + 1) *
        lucasU ⟨p, -1, by decide⟩ n
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 2.4.
Uses 3

TheoremShifted Catalan convolution with Lucas sequences#●ready to prove

Let \(p\) be an integer, let \(U_n\) and \(V_n\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\) and both satisfy \(X_{j+2} = p X_{j+1} + X_j\), and let \(C_j\) be the \(j\)-th Catalan number. Then for all natural numbers \(n\) and \(m\) with \(m \ge 1\), in the rationals,

\[\sum_{k=0}^{n} C_{n-k+m} C_{k+m} U_k V_{n-k} = 2 \binom{2m-1}{m} \frac{(n+m+2)(n+1)}{(n+2m+1)(n+2m+2)} C_{n+m+1} U_n.\]

The paper states \(m \in \mathbb{N}_0\); at \(m = 0\) the printed right side is twice the sum (\(p = 1\), \(n = 1\): the sum is \(2\), the printed right side is \(4\)), so \(m = 0\) is excluded.

public theorem_wanted lucasU_lucasV_sum_catalan_shifted (p : ℤ) (n m : ℕ) (hm : 0 < m) :
    ∑ k ∈ Finset.range (n + 1),
        (catalan (n - k + m) * catalan (k + m) : ℚ) *
          lucasU ⟨p, -1, by decide⟩ k * lucasV ⟨p, -1, by decide⟩ (n - k) =
      2 * ((2 * m - 1).choose m : ℚ) *
          (((n + m + 2) * (n + 1) : ℚ) / ((n + 2 * m + 1) * (n + 2 * m + 2))) *
        catalan (n + m + 1) * lucasU ⟨p, -1, by decide⟩ n
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 2.5.
Uses 3

TheoremConvolution of U and V with shifted binomial weights#●ready to prove

Let \(p\) be an integer and let \(U\) and \(V\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so that \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\), and both satisfy \(X_{j+2} = p X_{j+1} + X_j\). Then for every natural number \(n\)

\[\sum_{k=0}^{n} \binom{2n}{k} \binom{2n}{n+k} U_k V_{n-k} = \binom{4n}{n} U_n.\]
public theorem_wanted lucasU_lucasV_sum_choose_two_n (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (n + 1),
        ((2 * n).choose k * (2 * n).choose (n + k) : ℤ) *
          lucasU ⟨p, -1, by decide⟩ k * lucasV ⟨p, -1, by decide⟩ (n - k) =
      ((4 * n).choose n : ℤ) * lucasU ⟨p, -1, by decide⟩ n
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 2.6.
Uses 3

TheoremAlternating sum of \(\binom{2n}{k} \binom{4n}{2k}\) with Lucas sequences#●ready to prove

Let \(p\) be an integer and let \(U_n\) and \(V_n\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\) and both satisfy \(X_{j+2} = p X_{j+1} + X_j\). Then for every natural number \(n\), in the rationals,

\[\sum_{k=0}^{2n} (-1)^k \binom{2n}{k} \binom{4n}{2k} U_k V_{2n-k} = (-1)^n \frac{\binom{6n}{2n} \binom{2n}{n}}{\binom{3n}{n}} U_{2n}.\]
public theorem_wanted lucasU_lucasV_sum_choose_four_n_even_alternating (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (2 * n + 1),
        ((2 * n).choose k * (4 * n).choose (2 * k) : ℚ) * (-1) ^ k *
          lucasU ⟨p, -1, by decide⟩ k * lucasV ⟨p, -1, by decide⟩ (2 * n - k) =
      (-1) ^ n * ((6 * n).choose (2 * n) * (2 * n).choose n : ℚ) / (3 * n).choose n *
        lucasU ⟨p, -1, by decide⟩ (2 * n)
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 2.7.
Uses 3

TheoremAlternating triple-binomial convolution of U and V#●ready to prove

Let \(p\) be an integer and let \(U\) and \(V\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so that \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\), and both satisfy \(X_{j+2} = p X_{j+1} + X_j\). Then for every natural number \(n\)

\[\sum_{k=0}^{2n} (-1)^k \binom{2n}{k} \binom{4n}{k} \binom{4n}{2n+k} U_k V_{2n-k} = (-1)^n \binom{2n}{n} \binom{5n}{2n} U_{2n}.\]
public theorem_wanted lucasU_lucasV_sum_choose_triple_alternating (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (2 * n + 1),
        ((2 * n).choose k * (4 * n).choose k * (4 * n).choose (2 * n + k) : ℤ) * (-1) ^ k *
          lucasU ⟨p, -1, by decide⟩ k * lucasV ⟨p, -1, by decide⟩ (2 * n - k) =
      (-1) ^ n * ((2 * n).choose n * (5 * n).choose (2 * n) : ℤ) *
        lucasU ⟨p, -1, by decide⟩ (2 * n)
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 2.8.
Uses 3

TheoremAlternating sum with reciprocal binomial coefficients and Lucas sequences#●ready to prove

Let \(p\) be an integer and let \(U_n\) and \(V_n\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\) and both satisfy \(X_{j+2} = p X_{j+1} + X_j\). Then for every natural number \(n\), in the rationals,

\[\sum_{k=0}^{2n} (-1)^k \frac{\binom{2n}{k}}{\binom{4n}{k} \binom{4n}{2n+k}} U_k V_{2n-k} = \frac{4n+1}{3n+1} \cdot \frac{1}{\binom{4n}{2n}} U_{2n}.\]
public theorem_wanted lucasU_lucasV_sum_choose_inv_alternating (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (2 * n + 1),
        ((2 * n).choose k : ℚ) / ((4 * n).choose k * (4 * n).choose (2 * n + k)) * (-1) ^ k *
          lucasU ⟨p, -1, by decide⟩ k * lucasV ⟨p, -1, by decide⟩ (2 * n - k) =
      (4 * n + 1 : ℚ) / (3 * n + 1) / (4 * n).choose (2 * n) * lucasU ⟨p, -1, by decide⟩ (2 * n)
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 2.9.
Uses 3

TheoremSum of \(\binom{n}{k} \binom{2n}{n+k} \binom{n+k}{k}\) with Lucas sequences#●ready to prove

Let \(p\) be an integer and let \(U_n\) and \(V_n\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\) and both satisfy \(X_{j+2} = p X_{j+1} + X_j\). Then for every natural number \(n\),

\[\sum_{k=0}^{n} \binom{n}{k} \binom{2n}{n+k} \binom{n+k}{k} U_k V_{n-k} = \binom{2n}{n}^2 U_n.\]
public theorem_wanted lucasU_lucasV_sum_choose_n_plus_k (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (n + 1),
        (n.choose k * (2 * n).choose (n + k) * (n + k).choose k : ℤ) *
          lucasU ⟨p, -1, by decide⟩ k * lucasV ⟨p, -1, by decide⟩ (n - k) =
      ((2 * n).choose n : ℤ) ^ 2 * lucasU ⟨p, -1, by decide⟩ n
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 2.10.
Uses 3

TheoremBinomial-Catalan sum with Lucas sequences#●ready to prove

Let \(p\) be an integer, let \(U_n\) and \(V_n\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\) and both satisfy \(X_{j+2} = p X_{j+1} + X_j\), and let \(C_j\) be the \(j\)-th Catalan number. Then for every natural number \(n\),

\[\sum_{k=0}^{n} \binom{2n}{2k} C_k C_{n-k} U_{n-k} V_k = C_n C_{n+1} U_n.\]
public theorem_wanted lucasU_lucasV_sum_choose_catalan (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (n + 1),
        ((2 * n).choose (2 * k) * catalan k * catalan (n - k) : ℤ) *
          lucasU ⟨p, -1, by decide⟩ (n - k) * lucasV ⟨p, -1, by decide⟩ k =
      (catalan n * catalan (n + 1) : ℤ) * lucasU ⟨p, -1, by decide⟩ n
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 2.11.
Uses 3

TheoremAlternating binomial-Catalan sum with Lucas sequences#●ready to prove

Let \(p\) be an integer, let \(U_n\) and \(V_n\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\) and both satisfy \(X_{j+2} = p X_{j+1} + X_j\), and let \(C_j\) be the \(j\)-th Catalan number. Then for every natural number \(n\),

\[\sum_{k=0}^{2n} (-1)^k \binom{4n}{2k} C_k C_{2n-k} U_{2n-k} V_k = (-1)^n C_n C_{2n} U_{2n}.\]
public theorem_wanted lucasU_lucasV_sum_choose_catalan_alternating (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (2 * n + 1),
        (-1) ^ k * ((4 * n).choose (2 * k) * catalan k * catalan (2 * n - k) : ℤ) *
          lucasU ⟨p, -1, by decide⟩ (2 * n - k) * lucasV ⟨p, -1, by decide⟩ k =
      (-1) ^ n * (catalan n * catalan (2 * n) : ℤ) * lucasU ⟨p, -1, by decide⟩ (2 * n)
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 2.12.
Uses 3

TheoremAlternating two-binomial Catalan convolution of U and V#●ready to prove

Let \(p\) be an integer and let \(U\) and \(V\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so that \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\), and both satisfy \(X_{j+2} = p X_{j+1} + X_j\). Writing \(C_j\) for the \(j\)-th Catalan number, for every natural number \(n\)

\[\sum_{k=0}^{2n} (-1)^k \binom{2n}{k} \binom{4n}{2k} C_k C_{2n-k} U_{2n-k} V_k = (-1)^n \binom{3n+1}{n} C_n C_{2n} U_{2n}.\]
public theorem_wanted lucasU_lucasV_sum_choose_sq_catalan_alternating (p : ℤ) (n : ℕ) :
    ∑ k ∈ Finset.range (2 * n + 1),
        (-1) ^ k *
            ((2 * n).choose k * (4 * n).choose (2 * k) * catalan k * catalan (2 * n - k) : ℤ) *
          lucasU ⟨p, -1, by decide⟩ (2 * n - k) * lucasV ⟨p, -1, by decide⟩ k =
      (-1) ^ n * ((3 * n + 1).choose n * catalan n * catalan (2 * n) : ℤ) *
        lucasU ⟨p, -1, by decide⟩ (2 * n)
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 2.13.
Uses 3

TheoremBinomial sums with signs \((-1)^{\binom{k}{2}}\) and Lucas sequences#●ready to prove

Let \(p\) be an integer and let \(U_n\) and \(V_n\) be the Lucas sequences \(U(p, -1)\) and \(V(p, -1)\), so \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\), \(V_1 = p\) and both satisfy \(X_{j+2} = p X_{j+1} + X_j\). Then for every natural number \(n\),

\[\sum_{k=0}^{4n+1} (-1)^{\binom{k}{2}} \binom{4n+1}{k} U_k V_{4n+1-k} = (-1)^n 2^{2n+1} U_{4n+1}\]

and

\[\sum_{k=0}^{4n+1} (-1)^{\binom{k}{2}} \binom{4n+1}{k} V_k U_{4n+1-k} = (-1)^n 2^{2n+1} U_{4n+1}.\]
public theorem_wanted lucasU_lucasV_sum_choose_sign_choose_two (p : ℤ) (n : ℕ) :
    (∑ k ∈ Finset.range (4 * n + 2),
        ((4 * n + 1).choose k : ℤ) * (-1) ^ k.choose 2 *
          lucasU ⟨p, -1, by decide⟩ k * lucasV ⟨p, -1, by decide⟩ (4 * n + 1 - k) =
      (-1) ^ n * 2 ^ (2 * n + 1) * lucasU ⟨p, -1, by decide⟩ (4 * n + 1)) ∧
    ∑ k ∈ Finset.range (4 * n + 2),
        ((4 * n + 1).choose k : ℤ) * (-1) ^ k.choose 2 *
          lucasV ⟨p, -1, by decide⟩ k * lucasU ⟨p, -1, by decide⟩ (4 * n + 1 - k) =
      (-1) ^ n * 2 ^ (2 * n + 1) * lucasU ⟨p, -1, by decide⟩ (4 * n + 1)
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 3.1.
Uses 3

TheoremTwo-sided binomial sums of products of Lucas sequences#●ready to prove

Let \(p\) be an integer, let \(U\) and \(V\) be the two-sided second-order sequences for \(p\) with \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\) and \(V_1 = p\), both satisfying \(X_{j+1} = p X_j + X_{j-1}\) for every integer \(j\), and put \(\Delta = p^2 + 4\). Then for every natural number \(n > 0\) and all integers \(m\) and \(r\),

\[\sum_{k=-n}^{n} \binom{2n}{n+k} U_{k+m} U_{k+r} = \Delta^{n-1} V_{m+r},\]
\[\sum_{k=-n}^{n} \binom{2n}{n+k} V_{k+m} V_{k+r} = \Delta^{n} V_{m+r},\]
\[\sum_{k=-n}^{n} \binom{2n}{n+k} U_{k+m} V_{k+r} = \Delta^{n} U_{m+r},\]
\[\sum_{k=-n}^{n} \binom{2n}{n+k} V_{k+m} U_{k+r} = \Delta^{n} U_{m+r}.\]
public theorem_wanted secondOrderBiSeq_sum_central_choose_mul (p : ℤ) (n : ℕ) (hn : 0 < n)
    (m r : ℤ) :
    (∑ k ∈ Finset.Icc (-(n : ℤ)) n, ((2 * n).choose (n + k).toNat : ℤ) *
        secondOrderBiSeq p 0 1 (k + m) * secondOrderBiSeq p 0 1 (k + r) =
      (p ^ 2 + 4) ^ (n - 1) * secondOrderBiSeq p 2 p (m + r)) ∧
    (∑ k ∈ Finset.Icc (-(n : ℤ)) n, ((2 * n).choose (n + k).toNat : ℤ) *
        secondOrderBiSeq p 2 p (k + m) * secondOrderBiSeq p 2 p (k + r) =
      (p ^ 2 + 4) ^ n * secondOrderBiSeq p 2 p (m + r)) ∧
    (∑ k ∈ Finset.Icc (-(n : ℤ)) n, ((2 * n).choose (n + k).toNat : ℤ) *
        secondOrderBiSeq p 0 1 (k + m) * secondOrderBiSeq p 2 p (k + r) =
      (p ^ 2 + 4) ^ n * secondOrderBiSeq p 0 1 (m + r)) ∧
    ∑ k ∈ Finset.Icc (-(n : ℤ)) n, ((2 * n).choose (n + k).toNat : ℤ) *
        secondOrderBiSeq p 2 p (k + m) * secondOrderBiSeq p 0 1 (k + r) =
      (p ^ 2 + 4) ^ n * secondOrderBiSeq p 0 1 (m + r)
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 3.2.
Uses 1

TheoremWeighted two-sided binomial sums of products of Lucas sequences#●ready to prove

Let \(p\) be an integer, let \(U\) and \(V\) be the two-sided second-order sequences for \(p\) with \(U_0 = 0\), \(U_1 = 1\), \(V_0 = 2\) and \(V_1 = p\), both satisfying \(X_{j+1} = p X_j + X_{j-1}\) for every integer \(j\), and put \(\Delta = p^2 + 4\). Then for every natural number \(n > 0\) and all integers \(m\) and \(r\),

\[\sum_{k=-n}^{n} k \binom{2n}{n+k} U_{k+m} U_{k+r} = n p \Delta^{n-1} U_{m+r},\]
\[\sum_{k=-n}^{n} k \binom{2n}{n+k} V_{k+m} V_{k+r} = n p \Delta^{n} U_{m+r},\]
\[\sum_{k=-n}^{n} k \binom{2n}{n+k} U_{k+m} V_{k+r} = n p \Delta^{n-1} V_{m+r},\]
\[\sum_{k=-n}^{n} k \binom{2n}{n+k} V_{k+m} U_{k+r} = n p \Delta^{n-1} V_{m+r}.\]
public theorem_wanted secondOrderBiSeq_sum_central_choose_mul_weighted (p : ℤ) (n : ℕ)
    (hn : 0 < n) (m r : ℤ) :
    (∑ k ∈ Finset.Icc (-(n : ℤ)) n, ((2 * n).choose (n + k).toNat : ℤ) *
        secondOrderBiSeq p 0 1 (k + m) * secondOrderBiSeq p 0 1 (k + r) * k =
      n * p * (p ^ 2 + 4) ^ (n - 1) * secondOrderBiSeq p 0 1 (m + r)) ∧
    (∑ k ∈ Finset.Icc (-(n : ℤ)) n, ((2 * n).choose (n + k).toNat : ℤ) *
        secondOrderBiSeq p 2 p (k + m) * secondOrderBiSeq p 2 p (k + r) * k =
      n * p * (p ^ 2 + 4) ^ n * secondOrderBiSeq p 0 1 (m + r)) ∧
    (∑ k ∈ Finset.Icc (-(n : ℤ)) n, ((2 * n).choose (n + k).toNat : ℤ) *
        secondOrderBiSeq p 0 1 (k + m) * secondOrderBiSeq p 2 p (k + r) * k =
      n * p * (p ^ 2 + 4) ^ (n - 1) * secondOrderBiSeq p 2 p (m + r)) ∧
    ∑ k ∈ Finset.Icc (-(n : ℤ)) n, ((2 * n).choose (n + k).toNat : ℤ) *
        secondOrderBiSeq p 2 p (k + m) * secondOrderBiSeq p 0 1 (k + r) * k =
      n * p * (p ^ 2 + 4) ^ (n - 1) * secondOrderBiSeq p 2 p (m + r)
Sources
  • John M. Campbell and Emrah Kılıç, Binomial Sums Involving Second-Order Linearly Recurrent Sequences, The Fibonacci Quarterly 62 (2024), Article 62-1/campbell10202023, Proposition 3.3.
Uses 1
16 ready to prove

The dependency explorer needs JavaScript and its graph data.

Download the complete graph data.