Uneingeschränkter Zugang

On Mersenne Numbers and their Bihyperbolic Generalizations

 und   
29. Aug. 2024

Zitieren
COVER HERUNTERLADEN

Introduction and preliminary results

Let n ≥ 0 be an integer. The nth Mersenne number Mn and the nth Mersenne–Lucas number Hn are defined recursively by Mn=3Mn12Mn2, for n 2 withM0=0,M1=1 {M_n} = 3{M_{n - 1}} - 2{M_{n - 2}},{\rm{ for }}n \ge 2{\rm{ with}}\,{M_0} = 0,{M_1} = 1 and Hn=3Hn12Hn2, for n 2 with H0 = 2, H1=3 {H_n} = 3{H_{n - 1}} - 2{H_{n - 2}},\,\,{\rm{ for }}\,\,n \ge 2{\rm{with}}\,\,{H_0}{\rm{ }} = {\rm{ 2}},{\rm{ }}{H_1} = 3 respectively. Note that Mersenne–Lucas numbers are also called as Fermat numbers. The Binet type formulas of these sequences have the form Mn =2n 1 and Hn = 2n + 1, so Hn = Mn + 2.

Mersenne sequence has been studied in many papers, see for example [2, 3, 6, 7, 9]. In the literature, we can find some generalizations of Mersenne numbers, see [4, 10]. In [8], Ochalik and Włoch introduced the generalized Mersenne numbers as follows. Let k ≥ 3 be a fixed integer. For any integer n ≥ 0 let M (k, n) be the nth generalized Mersenne number defined by the second order linear recurrence relation of the form M(k,n)=kM(k,n1)(k1)M(k,n2) M(k,n) = kM(k,n - 1) - (k - 1)M(k,n - 2) for n ≥ 2 with M (k, 0) = 0 and M (k, 1) = 1.

For n = 0, 1, 2, 3, 4, . . . the generalized Mersenne numbers are 0, 1, k, k2 − k + 1, k3 2k2 + 2k, . . .. Moreover, M(3, n) = Mn.

By analogy, we define the generalized Mersenne–Lucas numbers in the following way. Let k ≥ 3 be a fixed integer. For any integer n ≥ 0 let H(k, n) be the nth generalized Mersenne–Lucas number defined by H(k,n)=kH(k,n1)(k1)H(k,n2) H(k,n) = kH(k,n - 1) - (k - 1)H(k,n - 2) for n ≥ 2 with H(k, 0) = 2 and H (k, 1) = 3.

Then the first few terms of the generalized Mersenne–Lucas sequence are 2, 3, k + 2, k2 − k + 3, k3 2k2 + 2k + 2, . . .. It is easily seen that H(3, n) = Hn.

Proposition 1.1

Let k3 be a fixed integer. For any integer n0 we have H(k, n) = M(k, n) + 2.

Proof

(By induction on n.) If n = 0 then M0 = 0, H0 = 2. If n = 1 then M1 = 1, H1 = 3. Now assume that for any n ≥ 0, we have H(k, n) = M(k, n)+2 and H(k, n+1) = M(k, n+1)+2. We shall show that H(k, n+2) = M(k, n + 2) + 2. Applying the induction’s hypothesis we obtain H(k,n+2)=kH(k,n+1)(k1)H(k,n)=k(M(k,n+1)+2)(k1)(M(k,n)+2)=kM(k,n+1)(k1)M(k,n)+2=M(k,n+2)+2, \matrix{ {H(k,n + 2)} \hfill & = \hfill & {kH(k,n + 1) - (k - 1)H(k,n)} \hfill \cr {} \hfill & = \hfill & {k(M(k,n + 1) + 2) - (k - 1)(M(k,n) + 2)} \hfill \cr {} \hfill & = \hfill & {kM(k,n + 1) - (k - 1)M(k,n) + 2} \hfill \cr {} \hfill & = \hfill & {M(k,n + 2) + 2,} \hfill \cr } and by the induction’s rule the formula follows.

Some identities, properties, combinatorial interpretations and matrix generators of M(k, n) were given in [8] and [11]. In the next part of the paper we use the following results.

Theorem 1.2 ([8])

Let n ≥ 0, k ≥ 3 be integers. Then M(k,n)=1k2k1n1. M(k,n) = {1 \over {k - 2}}\left( {{{\left( {k - 1} \right)}^n} - 1} \right).

Theorem 1.3 ([8])

Let n ≥ 0, k ≥ 3 be integers. Then M(k,n+1)=Mk,n=k1n. M(k,n + 1) = - M\left( {k,n} \right) = {\left( {k - 1} \right)^n}.

Theorem 1.4 ([11])

Let n ≥ 0, k ≥ 3 be integers. Then M(k,n+1)=k1Mk,n+1. M(k,n + 1) = \left( {k - 1} \right)M\left( {k,n} \right) + 1.

Using the fact that H(k, n) = M(k, n) + 2, we can give some properties of generalized Mersenne–Lucas numbers.

Corollary 1.5

Let n ≥ 0, k ≥ 3 be integers. Then H(k,n)=(k1)n+2k5k2, H(k,n) = {{{{(k - 1)}^n} + 2k - 5} \over {k - 2}}, H(k,n+1)H(k,n)=(k1)n H(k,n + 1) - H(k,n) = {(k - 1)^n} and H(k,n+1)=(k1)H(k,n)2k+5. H(k,n + 1) = (k - 1)H(k,n) - 2k + 5.

The Mersenne numbers and their generalizations have applications also in the theory of hypercomplex numbers. In [5], Daşdemir and Bilgici introduced and studied Mersenne quaternions, Gaussian Mersenne numbers and generalized Mersenne quaternions. In [11], the authors considered the Mersenne hybrid numbers and generalized Mersenne hybrid numbers. In this paper, we use the Mersenne, Mersenne–Lucas numbers and their generalizations in the theory of bihyperbolic numbers.

Hyperbolic numbers are two dimensional number system. Hyperbolic imaginary unit, so-called unipotent, is an element h±1 such that h2 = 1. Bihyperbolic numbers are a generalization of hyperbolic numbers. Let ℍ2 be the set of bihyperbolic numbers ζ of the form ς=x0+x1j1+x2j2+x3j3, \varsigma = {x_0} + {x_1}{j_1} + {x_2}{j_2} + {x_3}{j_3}, where x0, x1, x2, x3 ∈ 𝕉 and j1, j2, j3 ∉ 𝕉 R are operators such that j12=j22=j32=1,j1j2=j2j1=j3,j1j3=j3j1=j2,j2j3=j3j2=j1. j_1^2 = j_2^2 = j_3^2 = 1,{j_1}{j_2} = {j_2}{j_1} = {j_3},{j_1}{j_3} = {j_3}{j_1} = {j_2},{j_2}{j_3} = {j_3}{j_2} = {j_1}.

From the above rules the multiplication of bihyperbolic numbers can be made analogously to the multiplication of algebraic expressions. The addition and the subtraction of bihyperbolic numbers is done by adding and subtracting corresponding terms and hence their coefficients. The addition and multiplication on ℍ2 are commutative and associative, (ℍ2, +, ·) is a commutative ring. For the algebraic properties of bihyperbolic numbers, see [1].

Let n ≥ 0 be an integer. The nth bihyperbolic Mersenne number BhMn and the nth bihyperbolic Mersenne–Lucas number BhHn are defined by BhMn=Mn+Mn+1j1+Mn+2j2+Mn+3j3, Bh{M_n} = {M_n} + {M_{n + 1}}{j_1} + {M_{n + 2}}{j_2} + {M_{n + 3}}{j_3}, BhHn=Hn+Hn+1j1+Hn+2j2+Hn+3j3, Bh{H_n} = {H_n} + {H_{n + 1}}{j_1} + {H_{n + 2}}{j_2} + {H_{n + 3}}{j_3}, respectively, where Mn is the nth Mersenne number, Hn is the nth Mersenne–Lucas number and j1, j2, j3 are units which satisfy (1.9).

The nth generalized bihyperbolic Mersenne number BhMnk BhM_n^k we define in the following way BhMnk=M(k,n)+M(k,n+1)j1+M(k,n+2)j2+M(k,n+3)j3, BhM_n^k = M(k,n) + M(k,n + 1){j_1} + M(k,n + 2){j_2} + M(k,n + 3){j_3}, where M(k, n) denotes the nth generalized Mersenne number, defined by (1.1). By analogy, the nth bihyperbolic Mersenne–Lucas number BhHnk BhH_n^k is defined by BhHnk=H(k,n)+H(k,n+1)j1+H(k,n+2)j2+H(k,n+3)j3, BhH_n^k = H(k,n) + H(k,n + 1){j_1} + H(k,n + 2){j_2} + H(k,n + 3){j_3}, where H(k, n) denotes the nth generalized Mersenne–Lucas number, defined by (1.2). For k = 3 we have BhMn3=BhMn BhM_n^3 = Bh{M_n} and BhHn3=BhHn BhH_n^3 = Bh{H_n} .

Using the above definitions, we can write initial generalized bihyperbolic Mersenne numbers BhM0k=j1+kj2+(k2k+1)j3,BhM1k=1+kj1+(k2k+1)j2+(k32k2+2k)j3, \eqalign{& BhM_0^k = {j_1} + k{j_2} + ({k^2} - k + 1){j_3}, \cr & BhM_1^k = 1 + k{j_1} + ({k^2} - k + 1){j_2} + ({k^{_3}} - 2{k^2} + 2k){j_3}, \cr} generalized bihyperbolic Mersenne–Lucas numbers BhH0k=2+3j1+(k+2)j2+(k2k+3)j3,BhH1k=3+(k+2)j1+(k2k+3)j2+(k32k2+2k+2)j3, \eqalign{& BhH_0^k = 2 + 3{j_1} + (k + 2){j_2} + ({k^2} - k + 3){j_3}, \cr & BhH_1^k = 3 + (k + 2){j_1} + ({k^2} - k + 3){j_2} + ({k^3} - 2{k^2} + 2k + 2){j_3}, \cr} bihyperbolic Mersenne numbers BhM0=j1+3j2+7j3,BhM1=1+3j1+7j2+15j3, \eqalign{ & Bh{M_0} = {j_1} + 3{j_2} + 7{j_3}, \cr & Bh{M_1} = 1 + 3{j_1} + 7{j_2} + 15{j_3}, \cr} and bihyperbolic Mersenne–Lucas numbers BhH0=2+3j1+5j2+9j3,BhH1=3+5j1+9j2+17j3. \eqalign{& Bh{H_0} = 2 + 3{j_1} + 5{j_2} + 9{j_3}, \cr & Bh{H_1} = 3 + 5{j_1} + 9{j_2} + 17{j_3}. \cr}

Main results

In this section, we present some properties of the generalized bihyperbolic Mersenne and Mersenne–Lucas numbers.

Theorem 2.1

Let n ≥ 0, k ≥ 3 be integers. Then BhMn+2k=kBhMn+1k(k1)BhMnk, BhM_{n + 2}^k = kBhM_{n + 1}^k - (k - 1)BhM_n^k, where BhM0k BhM_0^k and BhM1k BhM_1^k are defined by (1.12).

Proof

By formulas (1.10) and (1.1) we get kBhMn+1k(k1)BhMnk =kM(k,n+1)+M(k,n+2)j1+M(k,n+3)j2+M(k,n+4)j3k1M(k,n)+M(k,n+1)j1+M(k,n+2)j2+M(k,n+3)j3=kMk,n+1k1Mk,n+kM(k,n+2)(k1)M(k,n+1)j1 +kM(k,n+3)(k1)M(k,n+2)j2 +kM(k,n+4)(k1)M(k,n+3)j3 =Mk,n+2+Mk,n+3j1+Mk,n+4j2+Mk,n+5j3 =BhMn+2k. \eqalign{ & kBhM_{n + 1}^k - (k - 1)BhM_n^k \cr & \, = k\left( {M(k,n + 1) + M(k,n + 2){j_1} + M(k,n + 3){j_2} + M(k,n + 4){j_3}} \right) \cr & \,\,\,\,\, - \left( {k - 1} \right)\left( {M(k,n) + M(k,n + 1){j_1} + M(k,n + 2){j_2} + M(k,n + 3){j_3}} \right) \cr & \, = kM\left( {k,n + 1} \right) - \left( {k - 1} \right)M\left( {k,n} \right) \cr & \,\,\,\,\, + \left( {kM(k,n + 2) - (k - 1)M(k,n + 1)} \right){j_1} \cr & \,\,\,\,\, + \left( {kM(k,n + 3) - (k - 1)M(k,n + 2)} \right){j_2} \cr & \,\,\,\,\, + \left( {kM(k,n + 4) - (k - 1)M(k,n + 3)} \right){j_3} \cr & \, = M\left( {k,n + 2} \right) + M\left( {k,n + 3} \right){j_1} + M\left( {k,n + 4} \right){j_2} + M\left( {k,n + 5} \right){j_3} \cr & \, = BhM_{n + 2}^k. \cr}

In the same way, using (1.11) and (1.2), we can prove the next theorem.

Theorem 2.2

Let n ≥ 0, k ≥ 3 be integers. Then BhHn+2k=kBhHn+1k(k1)BhHnk, BhH_{n + 2}^k = kBhH_{n + 1}^k - (k - 1)BhH_n^k, where BhH0k BhH_0^k and BhH1k BhH_1^k are defined by (1.13).

Theorem 2.3

Let n ≥ 0, k ≥ 3 be integers. Then BhMn+1k=k1BhMnk+1+j1+j2+j3, BhM_{n + 1}^k = \left( {k - 1} \right)BhM_n^k + 1 + {j_1} + {j_2} + {j_3}, where BhM0k BhM_0^k is defined by (1.12).

Proof

Using (1.10) and (1.5), we have BhMn+1k(k1)BhMnk =Mk,n+1+Mk,n+2j1+Mk,n+3j2+Mk,n+4j3 k1M(k,n)+M(k,n+1)j1+M(k,n+2)j2+M(k,n+3)j3=Mk,n+1k1Mk,n+Mk,n+2(k1)M(k,n+1))j1 +(M(k,n+3)(k1)M(k,n+2))j2 +(M(k,n+4)(k1)M(k,n+3))j3 =1+j1+j2+j3. \eqalign{ & BhM_{n + 1}^k - (k - 1)BhM_n^k \cr & = M\left( {k,n + 1} \right) + M\left( {k,n + 2} \right){j_1} + M\left( {k,n + 3} \right){j_2} + M\left( {k,n + 4} \right){j_3} \cr & \,\,\, - \left( {k - 1} \right)\left( {M(k,n) + M(k,n + 1){j_1} + M(k,n + 2){j_2} + M(k,n + 3){j_3}} \right) \cr & = M\left( {k,n + 1} \right) - \left( {k - 1} \right)M\left( {k,n} \right) + M\left( {k,n + 2} \right) - (k - 1)M(k,n + 1)){j_1} \cr & \,\,\, + (M(k,n + 3) - (k - 1)M(k,n + 2)){j_2} \cr & \,\,\, + (M(k,n + 4) - (k - 1)M(k,n + 3)){j_3} \cr & = 1 + {j_1} + {j_2} + {j_3}. \cr}

Theorem 2.4

Let n ≥ 0, k ≥ 3 be integers. Then BhHn+1k=(k1)BhHnk+(2k+5)(1+j1+j2+j3)2j14j26j3, BhH_{n + 1}^k = (k - 1)BhH_n^k + ( - 2k + 5)(1 + {j_1} + {j_2} + {j_3}) - 2{j_1} - 4{j_2} - 6{j_3}, where BhH0k BhH_0^k is defined by (1.13).

Proof

Using (1.11) and (1.8), we have BhHn+1k(k1)BhHnk =H(k,n+1)+H(k,n+2)j1+H(k,n+3)j2+H(k,n+4)j3 (k1)(H(k,n)+H(k,n+1)j1+H(k,n+2)j2+H(k,n+3)j3)=H(k,n+1)(k1)H(k,n)+H(k,n+2)(k1)H(k,n+1))j1 +(H(k,n+3)(k1)H(k,n+2))j2 +(H(k,n+4)(k1)H(k,n+3))j3 =2k+5+(2k+3)j1+(2k+1)j2+(2k1)j3 =(2k+5)(1+j1+j2+j3)2j14j26j3. \eqalign{ & BhH_{n + 1}^k - (k - 1)BhH_n^k \cr & \, = H(k,n + 1) + H(k,n + 2){j_1} + H(k,n + 3){j_2} + H(k,n + 4){j_3} \cr & \,\,\,\,\, - (k - 1)(H(k,n) + H(k,n + 1){j_1} + H(k,n + 2){j_2} + H(k,n + 3){j_3}) \cr & \, = H(k,n + 1) - (k - 1)H(k,n) + H(k,n + 2) - (k - 1)H(k,n + 1)){j_1} \cr & \,\,\,\, + (H(k,n + 3) - (k - 1)H(k,n + 2)){j_2} \cr & \,\,\, + (H(k,n + 4) - (k - 1)H(k,n + 3)){j_3} \cr & \, = - 2k + 5 + ( - 2k + 3){j_1} + ( - 2k + 1){j_2} + ( - 2k - 1){j_3} \cr & \, = ( - 2k + 5)(1 + {j_1} + {j_2} + {j_3}) - 2{j_1} - 4{j_2} - 6{j_3}. \cr}

Next theorems give the Binet formulas for the generalized bihyperbolic Mersenne and Mersenne–Lucas numbers.

Theorem 2.5

Let n ≥ 0, k ≥ 3 be integers. Then BhMnk=(k1)n1k2(1+j1+j2+j3)+(k1)n(j1+kj2+(k2k+1)j3). \eqalign{ & BhM_n^k = {{{{(k - 1)}^n} - 1} \over {k - 2}}(1 + j1 + j2 + j3) \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, + {(k - 1)^n}({j_1} + k{j_2} + ({k^2} - k + 1){j_3}). \cr}

Proof

Using (1.4), we have M(k, n + 1) = M(k, n) + (k − 1)n, hence M(k, n + 2) = M(k, n + 1) + (k − 1)n+1 = M(k, n) + (k − 1)n + (k − 1)n+1 and M(k, n + 3) = M(k, n) + (k − 1)n + (k − 1)n+1 + (k − 1)n+2. Thus BhMnk =M(k,n)+M(k,n+1)j1 +M(k,n+2)j2 +M(k,n+3)j3 =M(k,n)+(1+j1 +j2 +j3 ) +(k1) n j1 +(k1) n +(k1) n+1 j2 +((k1) n +(k1) n+1 +(k1) n+2 j3 =M(k,n)(1+j1 +j2 +j3 )+(k1) n (j1 +kj2 +(k2 k+1)j3 ). \eqalign{ & BhM_n^k = M(k,n) + M(k,n + 1){j_1} + M(k,n + 2){j_2} + M(k,n + 3){j_3} \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\, = M(k,n) + (1 + {j_1} + {j_2} + {j_3}) \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, + {(k - 1)^n}{j_1} + \left( {{{(k - 1)}^n} + {{(k - 1)}^{n + 1}}} \right){j_2} \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, + ({(k - 1)^n} + \left( {{{(k - 1)}^{n + 1}} + {{(k - 1)}^{n + 2}}} \right){j_3} \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\, = M(k,n)(1 + {j_1} + {j_2} + {j_3}) + {(k - 1)^n}({j_1} + k{j_2} + ({k^2} - k + 1){j_3}). \cr}

Putting M(k,n)=1k-2((k-1)n-1) M(k,n) = {1 \over {k - 2}}((k - 1)^n - 1) (see (1.3)), we obtain the desired formula.

Theorem 2.6

Let n ≥ 0, k ≥ 3 be integers. Then BhHnk=(k1)n+2k5k2(1+j1+j2+j3)+(k1)n(j1+kj2+(k2k+1)j3). \eqalign{ & BhH_n^k = {{{{(k - 1)}^n} + 2k - 5} \over {k - 2}}(1 + {j_1} + {j_2} + {j_3}) \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, + {(k - 1)^n}({j_1} + k{j_2} + ({k^2} - k + 1){j_3}). \cr}

Proof

Using (1.6), (1.7) and proceeding analogously as in the proof of the previous theorem we obtain the desired formula.

Corollary 2.7

Let n ≥ 0 be an integer. For k = 3 we have BhMn=(2n1)(1+j1+j2+j3)+2n(j1+3j2+7j3)=2n(1+2j1+4j2+8j3)(1+j1+j2+j3) \eqalign{ & Bh{M_n} = ({2^n} - 1)(1 + {j_1} + {j_2} + {j_3}) + {2^n}({j_1} + 3{j_2} + 7{j_3}) \cr & \,\,\,\,\,\,\,\,\,\,\, = {2^n}(1 + 2{j_1} + 4{j_2} + 8{j_3}) - (1 + {j_1} + {j_2} + {j_3}) \cr} and BhHn=(2n+1)(1+j1+j2+j3)+2n(j1+3j2+7j3)=2n(1+2j1+4j2+8j3)+(1+j1+j2+j3). \eqalign{ & Bh{H_n} = ({2^n} + 1)(1 + {j_1} + {j_2} + {j_3}) + {2^n}({j_1} + 3{j_2} + 7{j_3}) \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\, = {2^n}(1 + 2{j_1} + 4{j_2} + 8{j_3}) + (1 + {j_1} + {j_2} + {j_3}). \cr}

For simplicity of notation let A = 1 + j1 + j2 + j3. Using (1.3), (1.6) and (1.12), we can write (2.1) and (2.2) as BhMnk=AM(k,n)+(k1)nBhM0k BhM_n^k = A \cdot M(k,n) + {(k - 1)^n}BhM_0^k and BhHnk=AH(k,n)+(k1)nBhM0k, BhH_n^k = A \cdot H(k,n) + {(k - 1)^n}BhM_0^k, respectively.

Using the Binet formula (2.3) and identity (1.3), we can derive the Catalan identity for the generalized bihyperbolic Mersenne numbers.

Theorem 2.8

Let n ≥ 0, r ≥ 0, k ≥ 3 be integers such that nr. Then BhMn+rkBhMnrkBhMnk2=(k1)n1k2(k2+2)(k1)nr(1(k1)r) 2 (1+j1+j2+j3). \eqalign{ & BhM_{n + r}^k \cdot BhM_{n - r}^k - {\left( {BhM_n^k} \right)^2} \cr & \,\,\,\,\,\,\,\,\,\, = {{{{(k - 1)}^n} - 1} \over {k - 2}}({k^2} + 2){(k - 1)^{n - r}}{(1 - {(k - 1)^r})^2}(1 + {j_1} + {j_2} + {j_3}). \cr}

Proof

By formula (2.3) we get BhMn+rkBhMnrkBhMnk2=AM(k,n)+(k1)n+rBhM0kAM(k,n)+(k1)nrBhM0kAM(k,n)+(k1)nBhM0kAM(k,n)+(k1)nBhM0k=AM(k,n)BhM0kk1nr+AM(k,n)BhM0kk1n+r2AMk,nBhM0k(k1)n=AM(k,n)BhM0kk1nr1+(k1)2r2(k1)r=M(k,n)ABhM0kk1nr1(k1)r2. \eqalign{ & BhM_{n + r}^k \cdot BhM_{n - r}^k - {\left( {BhM_n^k} \right)^2} \cr & \, = \left( {A \cdot M(k,n) + {{(k - 1)}^{n + r}}BhM_0^k} \right)\,\left( {A \cdot M(k,n) + {{(k - 1)}^{n - r}}BhM_0^k} \right) \cr & \,\,\, - \left( {A \cdot M(k,n) + {{(k - 1)}^n}BhM_0^k} \right)\left( {A \cdot M(k,n) + {{(k - 1)}^n}BhM_0^k} \right) \cr & \, = A \cdot M(k,n) \cdot BhM_0^k \cdot {\left( {k - 1} \right)^{n - r}} + A \cdot M(k,n) \cdot BhM_0^k \cdot {\left( {k - 1} \right)^{n + r}} \cr & \,\,\, - 2A \cdot M\left( {k,n} \right) \cdot BhM_0^k \cdot {(k - 1)^n} \cr & \, = A \cdot M(k,n) \cdot BhM_0^k \cdot {\left( {k - 1} \right)^{n - r}}\left( {1 + {{(k - 1)}^{2r}} - 2{{(k - 1)}^r}} \right) \cr & \, = M(k,n) \cdot A \cdot BhM_0^k \cdot {\left( {k - 1} \right)^{n - r}}{\left( {1 - {{(k - 1)}^r}} \right)^2}. \cr}

Moreover, ABhM0k=1+j1+j2+j3j1+kj2+(k2k+1)j3=j1+kj2+k2k+1j3+1+kj3+k2k+1j2+j3+k+k2k+1j1+j2+kj1+k2k+1=(k2+2)(1+j1+j2+j3). \eqalign{ & A \cdot BhM_0^k = \left( {1 + {j_1} + {j_2} + {j_3}} \right)\left( {{j_1} + k{j_2} + ({k^2} - k + 1){j_3}} \right) \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = {j_1} + k{j_2} + \left( {{k^2} - k + 1} \right){j_3} + 1 + k{j_3} + \left( {{k^2} - k + 1} \right){j_2} \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, + {j_3} + k + \left( {{k^2} - k + 1} \right){j_1} + {j_2} + k{j_1} + \left( {{k^2} - k + 1} \right) \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = ({k^2} + 2)(1 + {j_1} + {j_2} + {j_3}). \cr}

Hence we get BhMn+rkBhMnrkBhMnk2=(k1)n1k2(k2+2)(k1)nr(1(k1)r)2(1+j1+j2+j3), \eqalign{ & BhM_{n + r}^k \cdot BhM_{n - r}^k - {\left( {BhM_n^k} \right)^2} \cr & = {{{{(k - 1)}^n} - 1} \over {k - 2}}({k^2} + 2){(k - 1)^{n - r}}{(1 - {(k - 1)^r})^2}(1 + {j_1} + {j_2} + {j_3}), \cr} which completes the proof.

In the same way, using (2.4) and (1.6), we obtain the Catalan identity for the generalized bihyperbolic Mersenne–Lucas numbers.

Theorem 2.9

Let n ≥ 0, r ≥ 0, k ≥ 3 be integers such that nr. Then BhHn+rkBhHnrkBhHnk2=(k1)n+2k5k2(k2+2)(k1)nr(1(k1)r)2(1+j1+j2+j3). \eqalign{ & BhH_{n + r}^k \cdot BhH_{n - r}^k - {\left( {BhH_n^k} \right)^2} \cr & = {{{{(k - 1)}^n} + 2k - 5} \over {k - 2}}({k^2} + 2){(k - 1)^{n - r}}{(1 - {(k - 1)^r})^2}(1 + {j_1} + {j_2} + {j_3}). \cr}

For r = 1 we obtain Cassini identities for the generalized bihyperbolic Mersenne and Mersenne–Lucas numbers.

Corollary 2.10

Let n ≥ 1, k ≥ 3 be integers. Then BhMn+1kBhMn1kBhMnk2=((k1)n1(k1)n1(k2)(k2+2)(1+j1+j2+j3). \eqalign{ & BhM_{n + 1}^k \cdot BhM_{n - 1}^k - {\left( {BhM_n^k} \right)^2} \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = ({(k - 1)^n} - 1{(k - 1)^{n - 1}}(k - 2)({k^2} + 2)(1 + {j_1} + {j_2} + {j_3}). \cr}

Corollary 2.11

Let n ≥ 1, k ≥ 3 be integers. Then BhHn+1kBhHn1kBhHnk2=((k1)n+2k5(k1)n1(k2)(k2+2)(1+j1+j2+j3). \eqalign{ & BhH_{n + 1}^k \cdot BhH_{n - 1}^k - {\left( {BhH_n^k} \right)^2} \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = ({(k - 1)^n} + 2k - 5{(k - 1)^{n - 1}}(k - 2)({k^2} + 2)(1 + {j_1} + {j_2} + {j_3}). \cr}

For k = 3 we obtain Catalan and Cassini identities for the bihyperbolic Mersenne and Mersenne–Lucas numbers.

Corollary 2.12

Let n1 be an integer. Then BhMn+rBhMnrBhMn2=11(2n1)(2nr2n+1+2n+r)(1+j1+j2+j3). \eqalign{ & Bh{M_{n + r}} \cdot Bh{M_{n - r}} - {\left( {Bh{M_n}} \right)^2} \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = 11({2^n} - 1)({2^{n - r}} - {2^{n + 1}} + {2^{n + r}})(1 + {j_1} + {j_2} + {j_3}). \cr}

Corollary 2.13

Let n ≥ 1 be an integer. Then BhHn+rBhHnrBhHn2=11(2n+1)(2nr2n+1+2n+r)(1+j1+j2+j3). \eqalign{ & Bh{H_{n + r}} \cdot Bh{H_{n - r}} - {\left( {Bh{H_n}} \right)^2} \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = 11({2^n} + 1)({2^{n - r}} - {2^{n + 1}} + {2^{n + r}})(1 + {j_1} + {j_2} + {j_3}). \cr}

Corollary 2.14

Let n ≥ 1 be an integer. Then BhMn+1BhMn1BhMn2=11(2n1)2n1(1+j1+j2+j3). Bh{M_{n + 1}} \cdot Bh{M_{n - 1}} - {\left( {Bh{M_n}} \right)^2}\, = 11({2^n} - 1){2^{n - 1}}(1 + {j_1} + {j_2} + {j_3}).

Corollary 2.15

Let n ≥ 1 be an integer. Then BhHn+1BhHn1BhHn2=11(2n+1)2n1(1+j1+j2+j3). Bh{H_{n + 1}} \cdot Bh{H_{n - 1}} - {\left( {Bh{H_n}} \right)^2}\, = 11({2^n} + 1){2^{n - 1}}(1 + {j_1} + {j_2} + {j_3}).

Now we give ordinary generating functions for the generalized bihyperbolic Mersenne and Mersenne–Lucas numbers.

Theorem 2.16

The generating function for the generalized bihyperbolic Mersenne number sequence BhMnk \left\{ {BhM_n^k} \right\} is G(t)=BhM0k+(BhM1kkBhM0k)t1kt+(k1)t2 G(t) = {{BhM_0^k + (BhM_1^k - kBhM_0^k)t} \over {1 - kt + (k - 1){t^2}}}

Proof

Assume that the generating function of the generalized bihyperbolic Mersenne number sequence BhMnk \left\{ {BhM_n^k} \right\} has the form G(t)=n=0BhMnktn G(t) = \sum\nolimits_{n = 0}^\infty {BhM_n^k{t^n}} Then (1kt+(k1)t2)G(t)=(1kt+(k1)t2)(BhM0k+BhM1kt+BhM2kt2+)=BhM0k+BhM1kt+BhM2kt2+kBhM0ktkBhM1kt2kBhM2kt3+(k1)BhM0kt2+(k1)BhM1kt3+(k1)BhM2kt4+=BhM0k+(BhM1kkBhM0k)t, \eqalign{ & (1 - kt + (k - 1){t^2})G(t) \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = (1 - kt + (k - 1){t^2})(BhM_0^k + BhM_1^kt + BhM_2^k{t^2} + \ldots ) \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = BhM_0^k + BhM_1^kt + BhM_2^k{t^2} + \ldots \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, - kBhM_0^kt - kBhM_1^k{t^2} - kBhM_2^k{t^3} - \ldots \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, + (k - 1)BhM_0^k{t^2} + (k - 1)BhM_1^k{t^3} + (k - 1)BhM_2^k{t^4} + \ldots \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = BhM_0^k + (BhM_1^k - kBhM_0^k)t, \cr} since BhMnk=kBhMn1kk1BhMn2k BhM_n^k = kBhM_{n - 1}^k - \left( {k - 1} \right)\,BhM_{n - 2}^k and the coefficients of tn for n ≥ 2 are equal to zero. Moreover, BhM0k=j1+kj2+(k2k+1)j3 BhM_0^k = {j_1} + k{j_2} + ({k^2} - k + 1){j_3} , BhM1kkBhM0k=1+k+1j2+k2+kj3 BhM_1^k - kBhM{_0^k} = 1 + \left({ - k + 1} \right){j_2} + \left({ - k^2} + k \right){j_ 3} .

Theorem 2.17

The generating function for the generalized bihyperbolic Mersenne-Lucas number sequence BhMnk \left\{ {BhM_n^k} \right\} is g(t)=BhH0k+(BhH1kkBhH0k)t1kt+(k1)t2. g(t) = {{BhH_0^k + (BhH_1^k - kBhH_0^k)t} \over {1 - kt + (k - 1){t^2}}}.

Proof

The proof of this theorem is similar to the proof of the previous theorem. Note only that BhH0k=2+3j1+k+2j2+k2k+3j3 BhH_0^k = 2 + 3{j_1} + \left( {k + 2} \right){j_2} + \left( {{k^2} - k + 3} \right){j_3} and BhH1kkBhH0k=32k+22kj1+33kj2+2kk2j3 BhH_1^k - kBhH_0^k = \left( {3 - 2k} \right) + \left( {2 - 2k} \right){j_1} + \left( {3 - 3k} \right)j{_2} + \left({2 - k - k^2} \right){j_3} .

Remark 2.18

The generating function γ(t) for the bihyperbolic Mersenne number sequence {BhMn} is γ(t)=BhM0+(BhM13BhM0)t13t+2t2, \gamma (t) = {{Bh{M_0} + (Bh{M_1} - 3Bh{M_0})t} \over {1 - 3t + 2{t^2}}}, where BhM0 = j1 + 3j2 + 7j3 and BhM1 3BhM0 = 1 2j2 6j3.

Remark 2.19

The generating function η(t) for the bihyperbolic Mersenne–Lucas number sequence {BhHn} is η(t)=BhH0+(BhH13BhH0)t13t+2t2, \eta (t) = {{Bh{H_0} + (Bh{H_1} - 3Bh{H_0})t} \over {1 - 3t + 2{t^2}}}, where BhH0 = 2+3j1 +5j2 +9j3 and BhH1 3BhH0 = 34j1 6j2 10j3.

At the end, we give the matrix representations of the defined bihyperbolic numbers.

Theorem 2.20

Let n ≥ 0, k ≥ 3 be integers. Then BhMn+2kBhMn+1kBhMn+1kBhMnk=BhM2kBhM1kBhM1kBhM0kk1(k1)0n. \left[ {\matrix{ {BhM_{n + 2}^k} & {BhM_{n + 1}^k} \cr {BhM_{n + 1}^k} & {BhM_n^k} \cr } } \right] = \left[ {\matrix{ {BhM_2^k} & {BhM_1^k} \cr {BhM_1^k} & {BhM_0^k} \cr } } \right] \cdot {\left[ {\matrix{ k \hfill & 1 \hfill \cr { - (k - 1)} \hfill & 0 \hfill \cr } } \right]^n}.

Proof

(By induction on n.) If n = 0 then assuming that the matrix to the power 0 is the identity matrix the result is obvious. Now suppose that for any n ≥ 0 holds BhMn+2kBhMn+1kBhMn+1kBhMnk=BhM2kBhM1kBhM1kBhM0kk1(k1)0n. \left[ {\matrix{ {BhM_{n + 2}^k} & {BhM_{n + 1}^k} \cr {BhM_{n + 1}^k} & {BhM_n^k} \cr } } \right] = \left[ {\matrix{ {BhM_2^k} & {BhM_1^k} \cr {BhM_1^k} & {BhM_0^k} \cr } } \right] \cdot {\left[ {\matrix{ k \hfill & 1 \hfill \cr { - (k - 1)} \hfill & 0 \hfill \cr } } \right]^n}.

We shall show that BhMn+3kBhMn+2kBhMn+2kBhMn+1k=BhM2kBhM1kBhM1kBhM0kk1(k1)0n+1. \left[ {\matrix{ {BhM_{n + 3}^k} & {BhM_{n + 2}^k} \cr {BhM_{n + 2}^k} & {BhM_{n + 1}^k} \cr } } \right] = \left[ {\matrix{ {BhM_2^k} & {BhM_1^k} \cr {BhM_1^k} & {BhM_0^k} \cr } } \right] \cdot {\left[ {\matrix{ k \hfill & 1 \hfill \cr { - (k - 1)} \hfill & 0 \hfill \cr } } \right]^{n + 1}}.

By simple calculations, using induction’s hypothesis we have [BhM2kBhM1kBhM1kBhM0k][k1-(k-1)0]n[k1-(k-1)0]=[BhMn+2kBhMn+1kBhMn+1kBhMnk][k1-(k-1)0]=[kBhMn+2k-(k-1)BhMn+1kBhMn+2kkBhMn+1k-(k-1)BhMnkBhMn+1k]=[BhMn+3kBhMn+2kBhMn+2kBhMn+1k], \eqalign{ & \left[ {\matrix{ {BhM_2^k} & {BhM_1^k} \cr {BhM_1^k} & {BhM_0^k} \cr } } \right] \cdot {\left[ {\matrix{ k \hfill & 1 \hfill \cr { - (k - 1)} \hfill & 0 \hfill \cr } } \right]^n} \cdot \left[ {\matrix{ k \hfill & 1 \hfill \cr { - (k - 1)} \hfill & 0 \hfill \cr } } \right] \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = \left[ {\matrix{ {BhM_{n + 2}^k} \hfill & {BhM_{n + 1}^k} \hfill \cr {BhM_{n + 1}^k} \hfill & {BhM_n^k} \hfill \cr } } \right] \cdot \left[ {\matrix{ k \hfill & 1 \hfill \cr { - (k - 1)} \hfill & 0 \hfill \cr } } \right] \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = \left[ {\matrix{ {k \cdot BhM_{n + 2}^k - (k - 1) \cdot BhM_{n + 1}^k} \hfill & {BhM_{n + 2}^k} \hfill \cr {k \cdot BhM_{n + 1}^k - (k - 1) \cdot BhM_n^k} \hfill & {BhM_{n + 1}^k} \hfill \cr } } \right] \cr & \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = \left[ {\matrix{ {BhM_{n + 3}^k} \hfill & {BhM_{n + 2}^k} \hfill \cr {BhM_{n + 2}^k} \hfill & {BhM_{n + 1}^k} \hfill \cr } } \right], \cr} which completes the proof.

Theorem 2.21

Let n ≥ 0, k ≥ 3 be integers. Then BhHn+2kBhHn+1kBhHn+1kBhHnk=BhH2kBhH1kBhH1kBhH0k.k1(k1)0n. \left[ {\matrix{ {BhH_{n + 2}^k} \hfill & {BhH_{n + 1}^k} \hfill \cr {BhH_{n + 1}^k} \hfill & {BhH_n^k} \hfill \cr } } \right] = \left[ {\matrix{ {BhH_2^k} \hfill & {BhH_1^k} \hfill \cr {BhH_1^k} \hfill & {BhH_0^k} \hfill \cr } } \right].{\left[ {\matrix{ k \hfill & 1 \hfill \cr { - (k - 1)} \hfill & 0 \hfill \cr } } \right]^n}.

Corollary 2.22

Let n ≥ 0 be an integer. Then BhMn+2BhMn+1BhMn+1BhMn=BhM2BhM1BhM1BhM0.3120n. \left[ {\matrix{ {Bh{M_{n + 2}}} \hfill & {Bh{M_{n + 1}}} \hfill \cr {Bh{M_{n + 1}}} \hfill & {Bh{M_n}} \hfill \cr } } \right] = \left[ {\matrix{ {Bh{M_2}} \hfill & {Bh{M_1}} \hfill \cr {Bh{M_1}} \hfill & {Bh{M_0}} \hfill \cr } } \right].{\left[ {\matrix{ 3 \hfill & 1 \hfill \cr { - 2} \hfill & 0 \hfill \cr } } \right]^n}.

Corollary 2.23

Let n ≥ 0 be an integer. Then BhHn+2BhHn+1BhHn+1BhHn=BhH2BhH1BhH1BhH0.3120n. \left[ {\matrix{ {Bh{H_{n + 2}}} \hfill & {Bh{H_{n + 1}}} \hfill \cr {Bh{H_{n + 1}}} \hfill & {Bh{H_n}} \hfill \cr } } \right] = \left[ {\matrix{ {Bh{H_2}} \hfill & {Bh{H_1}} \hfill \cr {Bh{H_1}} \hfill & {Bh{H_0}} \hfill \cr } } \right].{\left[ {\matrix{ 3 \hfill & 1 \hfill \cr { - 2} \hfill & 0 \hfill \cr } } \right]^n}.

Note that multiplication of bihyperbolic numbers is commutative and determinant properties can be used. For example, calculating determinants in Theorems 2.202.21 and Corollaries 2.222.23, we can also obtain Cassini identities. Using algebraic operations and matrix algebra could give many other interesting properties of these numbers.

Declarations

Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflict of Interest: The authors declare that they have no conflict of interest.

Sprache:
Englisch
Zeitrahmen der Veröffentlichung:
2 Hefte pro Jahr
Fachgebiete der Zeitschrift:
Mathematik, Mathematik, Allgemeines