Abstract
Let \(P\ge 3\) be an integer and let \((U_{n})\) and \((V_{n})\) denote generalized Fibonacci and Lucas sequences defined by \(U_{0}=0,U_{1}=1\); \( V_{0}=2,V_{1}=P,\) and \(U_{n+1}=PU_{n}-U_{n-1}\), \(V_{n+1}=PV_{n}-V_{n-1}\) for \(n\ge 1.\) In this study, when P is odd, we solve the equation \( U_{n}=wx^{2}+1\) for \(w=1,2,3,5,6,7,10.\) After then, we solve some Diophantine equations utilizing solutions of these equations.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Let P and Q be nonzero integers. Generalized Fibonacci sequence \((U_{n})\) and Lucas sequence \((V_{n})\) are defined by \(U_{0}(P,Q)=0,U_{1}(P,Q)=1\); \(V_{0}(P,Q)=2,V_{1}(P,Q)=P,\) and \(U_{n+1}(P,Q)=PU_{n}(P,Q)+QU_{n-1}(P,Q)\), \( V_{n+1}(P,Q)=PV_{n}(P,Q)+QV_{n-1}(P,Q)\) for \(n\ge 1.\) \(U_{n}(P,Q)\) and \( V_{n}(P,Q)\) are called n-th generalized Fibonacci number and n-th generalized Lucas number, respectively. Generalized Fibonacci and Lucas numbers for negative subscripts are defined as \( U_{-n}(P,Q)=-(-Q)^{-n}U_{n}(P,Q)\) and \(V_{-n}(P,Q)=(-Q)^{-n}V_{n}(P,Q),\) respectively.
Since
it will be assumed that \(P\ge 1.\) Moreover, we will assume that \( P^{2}+4Q>0. \) For \(P=Q=1,\) we have classical Fibonacci and Lucas sequences \( (F_{n})\) and \((L_{n}).\) For \(P=2\) and \(Q=1,\) we have Pell and Pell-Lucas sequences \((P_{n})\) and \((Q_{n}).\) For more information about generalized Fibonacci and Lucas sequences one can consult [1–4].
Generalized Fibonacci and Lucas numbers of the form \(kx^{2}\) have been investigated since 1962. In [5], the authors solved \( U_{n}=x^{2},V_{n}=x^{2},U_{n}=2x^{2},\) and \(V_{n}=2x^{2}\) for odd relatively prime integers P and Q. The reader can consult [6] or [7] for a brief discussion of the subject.
In [8], the authors showed that when \(a\ne 0\) and b are integers, then the equation \(U_{n}(P,\pm 1)=ax^{2}+b\) has only a finite number of solutions n. In [9], Keskin solved the equations \( V_{n}(P,-1)=wx^{2}+1\) and \(V_{n}(P,-1)=wx^{2}-1\) for \(w=1,2,3,6\) when P is odd. In [10], Karaatlı and Keskin solved the equations \( V_{n}(P,-1)=5x^{2}\pm 1\) and \(V_{n}(P,-1)=7x^{2}\pm 1\). Similar equations are tackled in [11] by using very different methods (see also [12–14]). In this study, we solve the equation \(U_{n}(P,-1)=wx^{2}+1\) for \(w=1,2,3,5,6,7,10.\)
We will use the Jacobi symbol throughout this study. Our method is elementary and used by Cohn, Ribenboim and McDaniel in [15] and [16] , respectively.
2 Preliminaries
From now on, instead of \(U_{n}(P,-1)\) and \(V_{n}(P,-1),\) we sometimes write \( U_{n}\) and \(V_{n},\) respectively. Moreover, we will assume that \(P\ge 3\).
The following lemmas can be proved by induction.
Lemma 2.1
If n is a positive integer, then \(U_{2n}\equiv n(-1)^{n+1}P\) \(( \mathrm{mod}P^{2})\) and \(U_{2n+1}\equiv (-1)^{n}\) \((\mathrm{mod}P^{2}).\)
Lemma 2.2
If n is a positive integer, then \(V_{2n}\equiv 2(-1)^{n}\) \(( \mathrm{mod}P^{2})\) and \(V_{2n+1}\equiv (-1)^{n}(2n+1)P\) \((\mathrm{mod}P^{2}).\)
The following theorems are given in [9].
Theorem 2.3
Let P be odd. If \(V_{n}=kx^{2}\) for some k|P with \(k>1,\) then \(n=1.\)
Theorem 2.4
Let P be odd. If \(V_{n}=2kx^{2}\) for some k|P with \(k>1,\) then \(n=3.\)
Theorem 2.5
Let P be odd. If \(U_{n}=kx^{2}\) for some k|P with \(k>1,\) then \( n=2\) or \(n=6\) and 3|P.
Theorem 2.6
Let P be odd. If k|P with \(k>1,\) then the equation \( U_{n}=2kx^{2}\) has no solutions.
Theorem 2.7
Let P be odd. If k|P with \(k>1,\) then the equation \( U_{n}=kx^{2}+1\) has only the solution \(n=1.\)
The following theorem is given in [10].
Theorem 2.8
Let P be odd. If \(V_{n}=7kx^{2}\) for some k|P with \(k>1,\) then \(n=1.\)
Now we give some known theorems from [5], which will be useful for solving the equation \(U_{n}=wx^{2}+1.\) We use a theorem from [17] while solving \(V_{n}=2x^{2}.\)
Theorem 2.9
Let P be odd. If \(V_{n}=x^{2}\) for some integer x, then \(n=1.\) If \(V_{n}=2x^{2}\) for some integer x, then \(n=3,P=3,27.\)
Theorem 2.10
Let P be odd. If \(U_{n}=x^{2}\) for some integer x, then \(n=1\) or \(n=2,\) \(P=\square \) or \(n=6,P=3.\) If \(U_{n}=2x^{2}\) for some integer x, then \(n=3.\)
The following lemma is a consequence of a theorem given in [18].
Lemma 2.11
All positive integer solutions of the equation \(3x^{2}-2y^{2}=1\) are given by \( (x,y)=(U_{n}(10,-1)-U_{n-1}(10,-1),U_{n}(10,-1)+U_{n-1}(10,-1)) \) with \( n\ge 1.\)
The proof of the following lemma is easy and will be omitted.
Lemma 2.12
All positive integer solutions of the equation \(x^{2}-7y^{2}=2\) are given by \((x,y)=\left( 3\left( U_{m+1}(16-1)-U_{m}(16,-1)\right) ,17U_{m}(16,-1)-U_{m-1}(16,-1)\right) \) with \(m\ge 0.\)
The following theorems are well known (see [19–22]).
Lemma 2.13
All positive integer solutions of the equation \( x^{2}-(P^{2}-4)y^{2}=4\) are given by \((x,y)=\left( V_{n}(P,-1),U_{n}(P,-1)\right) \) with \(n\ge 1.\)
Lemma 2.14
All positive integer solutions of the equation \( x^{2}-Pxy+y^{2}=1 \) are given by \((x,y)=\left( U_{n}(P,-1),U_{n-1}(P,-1)\right) \) with \(n\ge 2.\)
The following two theorems are given in [23].
Theorem 2.15
Let \(n\in \mathbb {N\cup }\left\{ 0\right\} ,\) m, \(r\in \mathbb {Z}\) and m be a nonzero integer. Then
Theorem 2.16
Let \(n\in \mathbb {N\cup }\left\{ 0\right\} \) and m, \(r\in \mathbb {Z}.\) Then
If \(n=2\cdot 2^{k}a+r\) with a odd, then we get
by (2.2).
When P is odd, since \(8|U_{3},\) using (2.1) we get
and therefore
From Lemma 2.1 and Lemma 2.2, it follows that if q|P with \( q>1, \) then
Now we give some identities concerning generalized Fibonacci and Lucas numbers:
Let \(m=2^{a}k,\) \(n=2^{b}l,\) k and l odd, \(a,b\ge 0,\) and \(d=(m,n).\) Then (see [24])
From (2.11) and Lemma 2.2, it follows that
An induction method shows that
and thus
and
for all \(k\ge 1.\)
Lemma 2.17
Let P be odd. Then
for all \(k\ge 1.\) Moreover, if \(3\not \mid P,\) then
for all \(k\ge 1.\)
Proof
If \(3\not \mid P,\) then \(P^{2}\equiv 1\) \((\mathrm{mod}3)\) and therefore \( V_{2}=P^{2}-2\equiv -1\) \((\mathrm{mod}3).\) An induction method shows that \( V_{2^{k}}\equiv -1\) \((\mathrm{mod}3)\) since \(V_{2^{k}}=\left( V_{2^{k-1}}\right) ^{2}-2\) by (2.12). Thus
Since \(V_{2}=P^{2}-2\equiv -1\) \((\mathrm{mod}P^{2}-1),\) it follows that \( V_{2^{k}}\equiv -1\) \((\mathrm{mod}P^{2}-1).\) Thus \(V_{2^{k}}\equiv -1\) \((\mathrm{ mod}P-1)\) and \(V_{2^{k}}\equiv -1\) \((\mathrm{mod}P+1).\) Let \(P-1=2^{t}a\) with a odd. Then we get
since \(\left( \frac{2}{V_{2^{k}}}\right) =1\) by (2.15). By using the fact that \(V_{2^{k}}\equiv -1\) \((\mathrm{mod}a),\) we get \(\left( \frac{P-1}{ V_{2^{k}}}\right) =\left( -1\right) ^{\frac{a-1}{2}}\left( \frac{-1}{a} \right) =1\) by (2.19). Similarly, it is seen that \(\left( \frac{P+1}{ V_{2^{k}}}\right) =1.\) \(\square \)
When P is odd, it can be shown that
and
for all \(k\ge 1.\)
3 Main theorems
From now on, we will assume that n is a positive integer and P is an odd integer.
Theorem 3.1
If \(U_{n}=2kx^{2}+1\) with k|P and \(k>1,\) then \(n=1\) or \(n=5.\)
Proof
Assume that \(U_{n}=2kx^{2}+1\) for some integer x. Then n is odd by Lemma 2.1. It is clear that \(n=1\) is a solution. Assume that \(n>1.\) Then we have \(n=2m+1\) with \(m\ge 1.\) Thus, we get \( U_{m}V_{m+1}=U_{2m+1}-1=2kx^{2}\) by (2.9). It can be seen that m is even by (2.6). Thus, \((U_{m},V_{m+1})=P\) by (2.13). Then it follows that
or
for some natural numbers a and b with \(k=k_{1}k_{2}.\) Thus, it is seen that
or
for some natural numbers u, v, s, t with u|P and v|P. Assume that (3.1) is satisfied. By using Theorems 2.4 and 2.10, it is seen that \(m=2.\) Therefore \(n=5.\) The identity (3.2) is impossible by Theorems 2.6 and 2.10.
Theorem 3.2
Let \(w=1,2,3,6.\) If \(U_{n}=wx^{2}+1\) for some integer x, then \( (w,n)=(1,1),(1,2),(2,1),(2,2),(3,1),(3,2),(6,1),(6,2),(6,5).\)
Proof
Assume that \(U_{n}=wx^{2}+1\) for some integer x. Let \(n>3\). Then \(n=4q+r\) for some \(q>0\) with \(0\le r\le 3\). Then \(n=2\cdot 2^{k}a+r\) with a odd and \(k\ge 1.\) Thus,
by (2.3). This shows that
Since \(\left( \dfrac{2}{V_{2^{k}}}\right) =1\), \(\left( \dfrac{-1}{V_{2^{k}}} \right) =-1,\) and \(\left( \dfrac{P+1}{V_{2^{k}}}\right) =1\) by (2.15), (2.16), and (2.17), respectively, we get
If \(w=1,2,\) then (3.3) is impossible. Let \(w=3,6.\) If \(3\not \mid P\), then again (3.3) is impossible since \(\left( \dfrac{3}{V_{2^{k}}} \right) =1\) by (2.18). Therefore \(n\le 3\) in case \(3\not \mid P\) and \( w=3,6.\) But \(n=3\) is not a solution in this case. If \(w=6\) and 3|P, then by Theorem 3.1, we get \(n=1\) or \(n=5.\) Thus, \(n=1,5\) for the case \(w=6\) and 3|P. If \(w=3\) and 3|P, then by Theorem 2.7, we get \(n=1.\)
\(\square \)
Theorem 3.3
If \(U_{n}=5x^{2}+1\) for some integer x, then \(n=1\) or \(n=2.\)
Proof
Assume that \(U_{n}=5x^{2}\) \(+1\) for some integer x. If 5|P, then by Theorem 2.7, \(n=1.\) Assume that \(5\not \mid P\). Let \(n>2\) and n be even. Now we divide the proof into two cases.
Case I. Let \(\ P^{2}\equiv 1\) \((\mathrm{mod}5).\) Since n is even, \( n=4q+r\) for some positive integer q with \(r=0,2.\) Thus, \(n=2\cdot 2^{k}a+r\) with a odd and \(k\ge 1.\) Then
by (2.3). This shows that
which is impossible since \(\left( \dfrac{-1}{V_{2^{k}}}\right) =-1\), \(\left( \dfrac{P+1}{V_{2^{k}}}\right) =1\), and \(\left( \dfrac{5}{V_{2^{k}}}\right) =1 \) by (2.16), (2.17), and (2.20), respectively.
Case II. Let \(\ P^{2}\equiv -1\) \((\mathrm{mod}5)\). We get \( 5x^{2}\equiv -1\) \((\mathrm{mod}P)\) since \(P|U_{n}\) when n is even. This shows that
which implies that \(P\equiv 3,7\) \((\mathrm{mod}8)\). Since n is even, we get \( n=6q+r\), \(r=0,2,4.\) Then \(5x^{2}+1\) \(\equiv U_{r}\) \((\mathrm{mod}8)\) by (2.5). If \(r=0,\) then we get \(5x^{2}\) \(\equiv -1\) \((\mathrm{mod}8),\) which is impossible. Let \(r=2\). Then \(5x^{2}\) \(+1\equiv U_{2}\) \((\mathrm{mod}8)\) by (2.5), which shows that \(5x^{2}\) \(+1\equiv P\) \((\mathrm{mod}8).\) But this is impossible since \(P\equiv 3,7\) \((\mathrm{mod}8).\) Let \(r=4\). Then \(n=12t+4\) or \(n=12t+10\) for some integer t. Let \(n=12t+10\). Then \(n=12q_{1}-2\) with \( q_{1}>0.\) Thus, \(n=2\cdot 2^{k}a-2\) with a odd and \(k\ge 1.\) Then it follows that
by (2.3), which implies that
This is impossible since \(\left( \dfrac{P-1}{V_{2^{k}}}\right) =1\) and \( \left( \dfrac{5}{V_{2^{k}}}\right) =-1\) by (2.17) and (2.20), respectively. Let \(n=12t+4.\) Since \(16|U_{6},\) we get \(5x^{2}+1=U_{n}\equiv U_{4}\) \((\mathrm{mod}16)\) by (2.1). A simple computation shows that \( 5x^{2}+1\equiv 1,5,6,14\) \((\mathrm{mod}16)\) and therefore \(U_{4}\equiv 1,5,6,14 \) \((\mathrm{mod}16)\). Moreover, we have \(5x^{2}+1=U_{n}\equiv U_{4}\equiv -P\) \( (\mathrm{mod}8)\) by (2.5). Using the fact that \(5x^{2}+1\equiv 1,5,6,14\) \((\mathrm{mod}8),\) we see that \(P\equiv 3,7\) \((\mathrm{mod}8).\) Since \(P\equiv 3,7\) \(\ (\mathrm{mod}8)\) and \(P^{3}-2P=U_{4}\equiv 1,5,6,14\) \((\mathrm{mod}16),\) it is seen that \(P\equiv 3,15\) \((\mathrm{mod}16).\) Let \(P\equiv 3\) \((\mathrm{mod} 16)\) and \(P\equiv 3\) \((\mathrm{mod}5).\) Since n is even, \(\ n=10q+r,\) \(r\in \{0,2,4,6,8\}.\) Using \(5|U_{5},\) we get \(5x^{2}+1=U_{n}\equiv U_{r}\) \((\mathrm{ mod}5)\) by (2.1). A simple computation shows that \(r=4.\) Since \( n=10q+4\) and \(n=12t+4,\) we get \(\ n=60k+4\) for some natural number k. Thus, by using (2.2), it is seen that
which implies that
since \(V_{5}=P(P^{4}-5P^{2}+5).\) This shows that
By using the facts that \((P^{3}-2P-1)/4\equiv 1\) \((\mathrm{mod}4),\) \( P^{4}-5P^{2}+5\equiv 1\) \((\mathrm{mod}5),\) \(P^{4}-5P^{2}+5\equiv 9\) \((\mathrm{mod }16),\) and \(-3P^{2}+P+5\equiv 13\) \((\mathrm{mod}16),\) we get
a contradiction. Let \(P\equiv 3\) \((\mathrm{mod}16)\) and \(P\equiv 2\) \((\mathrm{mod }5).\) Then \((P-1)/2\equiv 3\) \((\mathrm{mod}5)\) and \((P-1)/2\equiv 1\) \((\mathrm{ mod}8)\) and therefore
Moreover,
by (2.1). This implies that
This shows that
which is impossible by (3.4). Let \(P\equiv 15\) \((\mathrm{mod}16).\) Then \( (P^{2}-3)/2\equiv 3\) \((\mathrm{mod}5)\) and \((P^{2}-3)/2\equiv 7\) \((\mathrm{mod} 8).\) Moreover, we get
by (2.2). This shows that
and therefore
This is impossible since
Now assume that \(n>3\) and n is odd. Then \(n=2m+1\) with \(m>1.\) Therefore \( U_{2m+1}=5x^{2}+1,\) which implies that \(5x^{2}=U_{2m+1}-1=U_{m}V_{m+1}\) by ( 2.9). Let m be odd. Then \((U_{m},V_{m+1})=1\) by (2.13) and ( 2.8). Thus,
or
for some integers a and b. The identities (3.5) and (3.6) are impossible by (2.14) and Theorem 2.9, respectively. Let m be even. Then \((U_{m},V_{m+1})=P\) by (2.13). Thus,
or
for some integers a and b. The identities (3.7) and (3.8) are impossible by (2.14) and Theorem 2.3, respectively. Therefore \(n\le 3\). If \(n=3,\) we get \(P^{2}-1=U_{3}=5x^{2}+1,\) which implies that \( P^{2}\equiv 2\) \((\mathrm{mod}5).\) This is impossible. Thus, \(n=1\) or \(n=2.\) \(\square \)
Theorem 3.4
If \(U_{n}=7x^{2}+1\) for some integer x, then \(n=1,2,3.\)
Proof
Assume that \(U_{n}=7x^{2}\) \(+1\) for some integer x. If \(7|P,\ \) then by Theorem 2.7, \(n=1.\) Assume that \(7\not \mid P\). Let \(n>2\) and n be even. Then \(7x^{2}+1\equiv 0\) \((\mathrm{mod}P).\) This shows that \(\left( \dfrac{7}{P} \right) =\left( \dfrac{-1}{P}\right) ,\) which implies that \(\left( \dfrac{P}{ 7}\right) =1\). Therefore \(P\equiv 1,2,4\) \((\mathrm{mod}7).\) Now we distinguish three cases.
Case I. Let \(P\equiv 1\) \((\mathrm{mod}7).\) Since n is even, \(n=4q+r\) for some \(q>0\) with \(r=0,2.\) Thus, \(n=2\cdot 2^{k}a+r\) with a odd and \( k\ge 1.\) Then we get
by (2.3), which implies that
This is impossible since \(\left( \dfrac{-1}{V_{2^{k}}}\right) =-1\), \(\left( \dfrac{P+1}{V_{2^{k}}}\right) =1,\) and \(\left( \dfrac{7}{V_{2^{k}}}\right) =1 \) by (2.16), (2.17), and (2.21), respectively.
Case II. Let \(P\equiv 4\) \((\mathrm{mod}7).\) Then \(7|V_{2}\) and
by (2.2). This is impossible since \(7\not \mid (-1\pm U_{r})\) for \(r=0,2.\)
Case III. Let \(P\equiv 2\) \((\mathrm{mod}7).\) If \(n=4q+2,\) then \( n=4(q+1)-2=2\cdot 2^{k}a-2\) with a odd and \(k\ge 1.\) Thus, we get
by (2.3), which implies that
But this is impossible since \(\left( \dfrac{P-1}{V_{2^{k}}}\right) =1\) and \( \left( \dfrac{7}{V_{2^{k}}}\right) =-1\) by (2.17) and (2.21), respectively. Let \(n=4q.\) Then \(n=12t+r\) with \(r=0,4,8.\) Assume that \(n=12t.\) Since P is odd, we can write \(P^{2}-1=2^{m}a\) with a odd. Thus,
by (2.1), which implies that
This shows that \(\left( \dfrac{7}{a}\right) =\left( \dfrac{-1}{a}\right) \) and therefore \(\left( \dfrac{a}{7}\right) =1.\) Thus,
a contradiction. Assume that \(n=12t+4.\) Since \(16|U_{6},\) we get \( U_{n}\equiv U_{4}\) \((\mathrm{mod}16)\) by (2.1). This shows that \( 7x^{2}+1\equiv P^{3}-2P\) \((\mathrm{mod}16).\) Since \(7x^{2}+1\equiv 0,1,8,13\) \(( \mathrm{mod}16)\), a simple computation shows that \(P\equiv 11,15\) \((\mathrm{mod} 16).\) Let \(P\equiv 11\) \((\mathrm{mod}16).\) Then
by (2.1), which shows that \(7x^{2}\equiv -2\) \((\mathrm{mod}P-1).\) Thus, we get \(\left( \dfrac{7}{(P-1)/2}\right) =\) \(\left( \dfrac{-2}{(P-1)/2} \right) \) and therefore \(\left( \dfrac{(P-1)/2}{7}\right) =\) \(\left( \dfrac{2 }{(P-1)/2}\right) .\) But this is impossible since \((P-1)/2\equiv 5\) \((\mathrm{ mod}8)\) and \((P-1)/2\equiv 4\) \((\mathrm{mod}7).\) Let \(P\equiv 15\) \((\mathrm{mod} 16).\) By using a similar argument, it is seen that \(7x^{2}\equiv P-1\) \(( \mathrm{mod}P^{2}-3).\) This shows that
Since \((P^{2}-3)/2\equiv 4\) \((\mathrm{mod}7)\), \((P^{2}-3)/2\equiv 7\) \((\mathrm{ mod}8),\) and \((P-1)/2\equiv 7\) \((\mathrm{mod}8),\) we get
a contradiction. Assume that \(n=12t+8.\) Then we can write \(n=12m-4.\) A simple computation shows that \(P\equiv 1,5\) \((\mathrm{mod}16)\) in this case. Let \(P\equiv 1\) \((\mathrm{mod}16).\) Then
which implies that \(7x^{2}\equiv -2\) \((\mathrm{mod}P+1).\) Thus, we get
Therefore by using the facts that \((P+1)/2\equiv 1\) \((\mathrm{mod}8)\) and \( (P+1)/2\equiv 5\) \((\mathrm{mod}7),\) we get
a contradiction. Let \(P\equiv 5\) \((\mathrm{mod}16).\) Then
by (2.2), which implies that \(7x^{2}\equiv -(P+1)\) \((\mathrm{mod} P^{2}-3).\) By using the facts that \((P^{2}-3)/2\equiv 4\) \((\mathrm{mod}7)\), \( (P^{2}-3)/2\equiv 3\) \((\mathrm{mod}8),\) and \((P+1)/2\equiv 3\) \((\mathrm{mod}8),\) we get
a contradiction. Thus, we conclude that \(n\le 2.\) Now assume that n is odd. Then \(n=2m+1\) with \(m\ge 0.\) Thus, \(U_{2m+1}=7x^{2}+1,\) which implies that \(7x^{2}=U_{2m+1}-1=U_{m}V_{m+1}\) by (2.9). Let m be odd. Then \((U_{m},V_{m+1})=1\) by (2.13) and (2.8). Thus,
or
for some integers a and b. Assume that (3.9) is satisfied. Then by Theorem 2.10, we get \(m=1\) and therefore \(n=3.\) The identity (3.10) is impossible by Theorem 2.9. Let m be even. Then \( (U_{m},V_{m+1})=P\) by (2.13). This implies that
or
for some integers a and b. By using Theorems 2.3 and 2.8, we have in both cases that \(m+1=1\) and therefore \(n=1.\) Consequently, we have \( n=1,2,3.\) \(\square \)
Theorem 3.5
If \(U_{n}=10x^{2}+1\) for some integer x, then \(n=1,2.\)
Proof
If 5|P, then by Theorem 3.1, \(n=1\) or \(n=5.\) Assume that \(5\not \mid P\) . Let \(n>2\) and n be even. Then \(10x^{2}+1\equiv 0\) \((\mathrm{mod}P)\) since \( P|U_{n}\) when n is even. Therefore
If \(P\equiv \pm 1\) \((\mathrm{mod}5),\) then \(P\equiv 1,3\) \((\mathrm{mod}8).\) If \( P\equiv \pm 2\) \((\mathrm{mod}5),\) then \(P\equiv 5,7\) \((\mathrm{mod}8).\) The remainder of the proof is split into two cases.
Case I. Let \(P\equiv \pm 1\) \((\mathrm{mod}5).\) Since n is even, we get \(n=4q+r\) for some positive integer q with \(r=0,2.\) Thus, \(n=2\cdot 2^{k}a+r\) with a odd and \(k\ge 1.\) Then
by (2.3). This shows that
which is impossible since \(\left( \dfrac{2}{V_{2^{k}}}\right) =1\), \(\left( \dfrac{-1}{V_{2^{k}}}\right) =-1\), \(\left( \dfrac{P+1}{V_{2^{k}}}\right) =1,\) and \(\left( \dfrac{5}{V_{2^{k}}}\right) =1\) by (2.15), (2.16), ( 2.17), and (2.20), respectively.
Case II. Let \(P\equiv \pm 2\) \((\mathrm{mod}5).\) Since n is even, we get \(n=6q+r\) with \(r=0,2,4.\) Then \(10x^{2}+1\) \(\equiv U_{r}\) \((\mathrm{mod}8)\) by (2.5). If \(r=0,\) then we get \(10x^{2}\) \(\equiv -1\) \((\mathrm{mod}8),\) which is impossible. Let \(r=2\). Then \(10x^{2}\) \(+1\equiv U_{2}\) \((\mathrm{mod} 8)\), which shows that \(10x^{2}\) \(+1\equiv P\) \((\mathrm{mod}8),\) which is impossible since \(P\equiv 5,7\) \((\mathrm{mod}8).\) Let \(r=4.\) Then either \( n=12t+10\) or \(n=12t+4\) for some nonnegative integer t. Assume that \( n=12t+10.\) Then \(n=12q_{1}-2\) with \(q_{1}>0.\) Thus, \(n=2\cdot 2^{k}a-2\) with a odd and \(k\ge 1.\) This shows that
by (2.3), which shows that
This is impossible since \(\left( \dfrac{2}{V_{2^{k}}}\right) =1\), \(\left( \dfrac{P-1}{V_{2^{k}}}\right) =1,\) and \(\left( \dfrac{5}{V_{2^{k}}}\right) =-1\) by (2.15), (2.17), and (2.20), respectively. Assume that \(n=12t+4.\) It can be seen that \(U_{n}=10x^{2}+1\) \(\equiv 1,9,11\) \(( \mathrm{mod}16).\) Moreover, we get \(U_{n}\equiv U_{4}\) \((\mathrm{mod}16)\) by ( 2.1) since \(16|U_{6}.\) A simple computation shows that \(P\equiv 7,13,15 \) \((\mathrm{mod}16)\) since \(P\equiv 5,7\) \((\mathrm{mod}8)\) and \(U_{4}=P^{3}-2P.\) Let \(P\equiv 7,15\) \((\mathrm{mod}16).\) Then \((P^{2}-3)/2\equiv 3\) \((\mathrm{mod} 5)\) and \((P^{2}-3)/2\equiv 7\) \((\mathrm{mod}8).\) Moreover, we get
by (2.2). This shows that
and therefore
Then we get
This is impossible since
Let \(P\equiv 13\) \((\mathrm{mod}16)\) and \(P\equiv 2\) \((\mathrm{mod}5).\) Then
and therefore
Moreover,
by (2.1). This implies that
This shows that
and therefore
which is impossible by (3.13). Let \(P\equiv 13\) \((\mathrm{mod}16)\) and \( P\equiv 3\) \((\mathrm{mod}5).\) Since n is even, \(\ n=10q+r\) with \(r\in \{0,2,4,6,8\}.\) Since \(5|U_{5}\) by (2.7), we get \(10x^{2}+1=U_{n} \equiv U_{r}\) \((\mathrm{mod}5)\) by (2.1). A simple computation shows that \(r=4.\) Since \(n=10q+4\) and \(n=12t+4,\) we get \(\ n=60k+4\) for some natural number k. Thus, by using (2.2), it is seen that
which implies that
since \(V_{5}=P(P^{4}-5P^{2}+5).\) This shows that
and therefore
Since \((P^{3}-2P-1)/2\equiv 5\) \((\mathrm{mod}8),\) \(P^{4}-5P^{2}+5\equiv 1\) \(( \mathrm{mod}5),\) \(P^{4}-5P^{2}+5\equiv 9\) \((\mathrm{mod}16),\) and \( -3P^{2}+P+5\equiv 7\) \((\mathrm{mod}16),\) we get
a contradiction. Now assume that \(n>1\) and n is odd. Then \(n=2m+1\) with \( m\ge 1.\) Therefore \(U_{2m+1}=10x^{2}+1,\) which implies that \( 10x^{2}=U_{2m+1}-1=U_{m}V_{m+1}\) by (2.9). Let m be odd. Then \( (U_{m},V_{m+1})=1\) by (2.13) and (2.8). Thus,
or
for some integers a and b. The identity (3.15) is impossible by Theorem 2.9. The identities (3.14) and (3.16) are impossible by (2.14), and (3.17) is impossible by Theorem . Let m be even. Then \((U_{m},V_{m+1})=P\) by (2.13). Thus,
or
for some integers a and b. The identities (3.18) and (3.20) are impossible by (2.14), and (3.19) is impossible by Theorem 2.3. Assume that (3.21) is satisfied. Then by Theorem 2.4, we get \(m=2\) and therefore \(n=5.\) Consequently, we have \(n=1,2,5\). But it can be seen that 5 is not a solution and therefore \(n=1,2.\) \(\square \)
By using MAGMA [25], it can be shown that the equation \( 2Px^{2}+1=U_{5}=P^{4}-3P^{2}+1\) has only the solution \(P=3.\) Therefore we can give the following corollary by using Theorem 3.1 and Lemmas 2.13 and 2.14.
Corollary 3.6
The equations \(x^{2}-(P^{2}-4)(2Py^{2}+1)^{2}=4\) and \( (2Px^{2}+1)^{2}-P(2Px^{2}+1)y+y^{2}=1\) have positive integer solutions only when \(P=3\). The solutions are given by \((x,y)=(123,3)\) and \((x,y)=(3,21),\) respectively.
Corollary 3.7
Let \(k=1,2,3,5,10.\) The equations \(x^{2}-(P^{2}-4)(ky^{2}+1)^{2}=4\) and \( (kx^{2}+1)^{2}-P(kx^{2}+1)y+y^{2}=1\) have positive integer solutions only when \(P=ka^{2}+1\) for some integer a.
Corollary 3.8
The equations \(x^{2}-(P^{2}-4)(6y^{2}+1)^{2}=4\) and \( (6x^{2}+1)^{2}-P(6x^{2}+1)y+y^{2}=1\) have positive integer solutions only when \(P=6a^{2}+1\) for some integer a or \(P=3(U_{m}(10,-1)+U_{m-1}(10,-1))\) for some \(m\ge 1\) and there is only one solution in each case.
Proof
In order to prove the corollary we must solve the equation \( 6x^{2}+1=U_{5}=P^{4}-3P^{2}+1.\) Since \(6x^{2}+1=P^{4}-3P^{2}+1,\) it is seen that \(P=3a\) and \(x=3b\) for some integers a and b. Then we get \( a^{2}(3a^{2}-1)=2b^{2},\) which implies that \(3a^{2}-1=2v^{2}.\) This shows that \(3a^{2}-2v^{2}=1.\) Thus, by Lemma 2.11, we get \(a=\) \( U_{m}(10,-1)-U_{m-1}(10,-1)\) for some \(m\ge 1.\) Since \(P=3a,\) the proof follows. \(\square \)
From Theorem 3.4 and Lemma 2.12, we can give the following corollary easily.
Corollary 3.9
The equations \(x^{2}-(P^{2}-4)(7y^{2}+1)^{2}=4\) and \( (7x^{2}+1)^{2}-P(7x^{2}+1)y+y^{2}=1\) have positive integer solutions only when \(P=7a^{2}+1\) for some integer a or \(P=3\left( U_{m+1}(16-1)-U_{m}(16,-1)\right) \) for some \(m\ge 1\) and there is only one solution in each case.
References
D. Kalman, R. Mena, The Fibonacci numbers-exposed. Math. Mag. 76, 167–181 (2003)
J.B. Muskat, Generalized Fibonacci and Lucas sequences and rootfinding methods. Math. Comput. 61, 365–372 (1993)
S. Rabinowitz, Algorithmic manipulation of Fibonacci identities. Appl Fibonacci Number 6, 389–408 (1996)
P. Ribenboim, My Numbers, My Friends (Springer, New York, 2000)
P. Ribenboim, W.L. McDaniel, The square terms in Lucas sequences. J. Number Theory 58, 104–123 (1996)
Z. Şiar, R. Keskin, The square terms in generalized Fibonacci sequence. Mathematika 60, 85–100 (2014)
R. Keskin, O. Karaatlı, Generalized Fibonacci and Lucas numbers of the form \(5x^{2}\). Int. J. Number. Theory 11(3), 931–944 (2015)
M.A. Alekseyev, S. Tengely, On integral points on biquadratic curves and near-multiples of squares in Lucas sequences. J. Integer Seq. 17, no. 6, Article ID 14.6.6, (2014)
R. Keskin, Generalized Fibonacci and Lucas numbers of the form \(wx^{2}\) and \(wx^{2}\pm 1\). Bull. Korean Math. Soc. 51, 1041–1054 (2014)
O. Karaatlı, R. Keskin, Generalized Lucas Numbers of the form \(5kx^{2}\) and \(7kx^{2}\). Bull. Korean Math. Soc. 52(5), 1467–1480 (2015)
M.A. Bennett, S. Dahmen, M. Mignotte, S. Siksek, Shifted powers in binary recurrence sequences. Math. Proc. Camb. Philos. Soc. 158(2), 305–329 (2015)
Y. Bugeaud, F. Luca, M. Mignotte, S. Siksek, Fibonacci numbers at most one away from a perfect power. Elem. Math. 63(2), 65–75 (2008)
Y. Bugeaud, F. Luca, M. Mignotte, S. Siksek, Almost powers in the Lucas sequence. J. Théor. Nombres Bordx. 20(3), 555–600 (2008)
Y. Bugeaud, M. Mignotte, S. Siksek, Classical and modular approaches to exponential Diophantine equations. I. Fibonacci and Lucas perfect powers. Ann. Math. 163(3), 969–1018 (2006)
J.H.E. Cohn, Squares in some recurrent sequences. Pac. J. Math. 41, 631–646 (1972)
P. Ribenboim, W. L. McDaniel, On Lucas sequence terms of the form \(kx^{2},\) number theory: proceedings of the Turku symposium on Number Theory in memory of Kustaa Inkeri (Turku, 1999), de Gruyter, Berlin, 293–303 (2001)
R.T. Bumby, The Diophantine equation \(3x^{4}-2y^{2}=1\). Math. Scand. 21, 144–148 (1967)
R. Keskin and Z. Şiar, Positive integer solutions of some Diophantine equations in terms of integer sequences (submitted)
J.P. Jones, Representation of solutions of Pell equations using Lucas sequences. Acta Acad. Paedagog. Agriensis Sect. Mat. 30, 75–86 (2003)
R. Keskin, Solutions of some quadratic Diophantine equations. Comput. Math. Appl. 60, 2225–2230 (2010)
W.L. McDaniel, Diophantine representation of Lucas sequences. Fibonacci Quart. 33, 58–63 (1995)
R. Melham, Conics which characterize certain Lucas sequences. Fibonacci Quart. 35, 248–251 (1997)
Z. Şiar, R. Keskin, Some new identities concerning generalized Fibonacci and Lucas numbers. Hacet. J. Math. Stat. 42(3), 211–222 (2013)
W.L. McDaniel, The g.c.d. in Lucas sequences and Lehmer number sequences. Fibonacci Quart. 29, 24–30 (1991)
W. Bosma, J. Cannon, C. Playoust, The MAGMA algebra system. I: the user language. J. Symb. Comput. 24(3–4), 235–265 (1997)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Keskin, R., Öğüt, Ü. Generalized Fibonacci numbers of the form \(wx^{2}+1\) . Period Math Hung 73, 165–178 (2016). https://doi.org/10.1007/s10998-016-0133-4
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10998-016-0133-4