Loading [MathJax]/jax/output/SVG/jax.js
Research article

Exact divisibility by powers of the integers in the Lucas sequences of the first and second kinds

  • Lucas sequences of the first and second kinds are, respectively, the integer sequences (Un)n0 and (Vn)n0 depending on parameters a,bZ and defined by the recurrence relations U0=0, U1=1, and Un=aUn1+bUn2 for n2, V0=2, V1=a, and Vn=aVn1+bVn2 for n2. In this article, we obtain exact divisibility results concerning Ukn and Vkn for all positive integers n and k. This and our previous article extend many results in the literature and complete a long investigation on this problem from 1970 to 2021.

    Citation: Kritkhajohn Onphaeng, Prapanpong Pongsriiam. Exact divisibility by powers of the integers in the Lucas sequences of the first and second kinds[J]. AIMS Mathematics, 2021, 6(11): 11733-11748. doi: 10.3934/math.2021682

    Related Papers:

    [1] Kritkhajohn Onphaeng, Prapanpong Pongsriiam . Exact divisibility by powers of the integers in the Lucas sequence of the first kind. AIMS Mathematics, 2020, 5(6): 6739-6748. doi: 10.3934/math.2020433
    [2] Tingting Du, Li Wang . On the power sums problem of bi-periodic Fibonacci and Lucas polynomials. AIMS Mathematics, 2024, 9(4): 7810-7818. doi: 10.3934/math.2024379
    [3] Hong Kang . The power sum of balancing polynomials and their divisible properties. AIMS Mathematics, 2024, 9(2): 2684-2694. doi: 10.3934/math.2024133
    [4] Faik Babadağ, Ali Atasoy . On hyper-dual vectors and angles with Pell, Pell-Lucas numbers. AIMS Mathematics, 2024, 9(11): 30655-30666. doi: 10.3934/math.20241480
    [5] Waleed Mohamed Abd-Elhameed, Amr Kamel Amin, Nasr Anwer Zeyada . Some new identities of a type of generalized numbers involving four parameters. AIMS Mathematics, 2022, 7(7): 12962-12980. doi: 10.3934/math.2022718
    [6] Yulu Feng, Min Qiu . Some results on p-adic valuations of Stirling numbers of the second kind. AIMS Mathematics, 2020, 5(5): 4168-4196. doi: 10.3934/math.2020267
    [7] Waleed Mohamed Abd-Elhameed, Omar Mazen Alqubori . New expressions for certain polynomials combining Fibonacci and Lucas polynomials. AIMS Mathematics, 2025, 10(2): 2930-2957. doi: 10.3934/math.2025136
    [8] Ümit Tokeşer, Tuğba Mert, Yakup Dündar . Some properties and Vajda theorems of split dual Fibonacci and split dual Lucas octonions. AIMS Mathematics, 2022, 7(5): 8645-8653. doi: 10.3934/math.2022483
    [9] Man Chen, Huaifeng Chen . On ideal matrices whose entries are the generalized $ k- $Horadam numbers. AIMS Mathematics, 2025, 10(2): 1981-1997. doi: 10.3934/math.2025093
    [10] Adikanda Behera, Prasanta Kumar Ray . Determinantal and permanental representations of convolved (u, v)-Lucas first kind p-polynomials. AIMS Mathematics, 2020, 5(3): 1843-1855. doi: 10.3934/math.2020123
  • Lucas sequences of the first and second kinds are, respectively, the integer sequences (Un)n0 and (Vn)n0 depending on parameters a,bZ and defined by the recurrence relations U0=0, U1=1, and Un=aUn1+bUn2 for n2, V0=2, V1=a, and Vn=aVn1+bVn2 for n2. In this article, we obtain exact divisibility results concerning Ukn and Vkn for all positive integers n and k. This and our previous article extend many results in the literature and complete a long investigation on this problem from 1970 to 2021.



    Throughout this article, let a and b be relatively prime integers and let (Un)n0 and (Vn)n0 be the Lucas sequences of the first and second kinds which are defined by the recurrence relations

    U0=0,U1=1,Un=aUn1+bUn2 for n2,
    V0=2,V1=a, and Vn=aVn1+bVn2 for n2.

    To avoid triviality, we also assume that b0 and α/β is not a root of unity where α and β are the roots of the characteristic polynomial x2axb. In particular, this implies that αβ, αβ, the discriminant D=a2+4b0, Un0, and Vn0 for all n1. If a=b=1, then (Un)n0 reduces to the sequence of Fibonacci numbers Fn; if a=6 and b=1, then (Un)n0 becomes the sequence of balancing numbers; if a=2 and b=1, then (Un)n0 is the sequence of Pell numbers; and many other famous integer sequences are just special cases of the Lucas sequences of the first and second kinds.

    The divisibility by powers of the Fibonacci numbers has attracted some attentions because it is used in Matijasevich's solution to Hilbert's 10th problem [5,6,7]. More precisely, Matijasevich show that

    F2nFnmif and only ifFnm. (1.1)

    From that point, Hoggatt and Bicknell-Johnson [4], Benjamin and Rouse [1], Seibert and Trojovský [19], Pongsriiam [15], Onphaeng and Pongsriiam [9,10], Panraksa and Tangboonduangjit [11], and Patra, Panda, and Khemaratchatakumthorn [12] have made some contributions on the extensions of (1.1). For more details about the timeline and the development of this problem, we refer the reader to the introduction of our previous article [8]. In fact, the most general results in this direction has recently been given by us [8] as follows.

    Theorem 1. [8,Theorem 10] Let k,m,nN, a,bZ, (a,b)=1, n2, and Uknm. Then

    (i) if a is odd and b is even, then Uk+1nUnm;

    (ii) if a is even and b is odd, then Uk+1nUnm;

    (iii) if a and b are odd and n3(mod6), then Uk+1nUnm;

    (iv) if a and b are odd, n3(mod6), and Uk+1n2m, then Uk+1nUnm;

    (v) if a and b are odd, n3(mod6), Uk+1n2m, and 2a2+3b, then Uk+1nUnm;

    (vi) if a and b are odd, n3(mod6), Uk+1n2m, and 4a2+3b, then Uk+t+1nUnm, where

    t=min({v2(U6)2}{ypk p is an odd prime factor of Un})andyp=vp(m)vp(Un)for each odd prime p dividing Un.

    Theorem 2. [8,Theorem 12] Let k,m,nN, a,bZ, (a,b)=1, n2, and Uk+1nUnm. Then

    (i) if a is odd and b is even, then Uknm;

    (ii) if a is even and b is odd, then Uknm;

    (iii) if a and b are odd and n3(mod6), then Uknm;

    (iv) if a and b are odd, n3(mod6), and 2a2+3b, then Uknm;

    (v) if a and b are odd, n3(mod6), 4a2+3b, and v2(m)k, then Uknm;

    (vi) if a and b are odd, n3(mod6), 4a2+3b, and v2(m)<k, then

     m is even, v2(m)k+1v2(a2+3b),and Uv2(m)n m.

    For other related and recent results on Fibonacci, Lucas, balancing, and Lucas-balancing numbers, see for example in [3,13,14,16,17,20] and references there in.

    In this article, we extend Theorems 1 and 2 to the case of Vn and the mix of Un and Vn. For example, we obtain in Theorem 18 that if a and m are even, b is odd, and Vk+1nUnm, then 2n implies Vmin(k,v2(m))nm; while 2n implies Vknm and the exponent k can be replaced by k+1 if and only if Vk+2n2Unm.

    In this section, we recall some definition and well known results, and give some useful lemmas for the reader's convenience. The order (or the rank) of appearance of nN in the Lucas sequence (Un)n0 is defined as the smallest positive integer m such that nUm and is denoted by τ(n). The exact divisibility mkn means that mkn and mk+1n. The letter p is always a prime. For nN, the p-adic valuation of n, denoted by vp(n) is the power of p in the prime factorization of n. We sometimes write the expression such as abc=d to mean that ab, bc, and c=d. For each xR, we write x to denote the largest integer less than or equal to x. So xx<x+1. We let D=a2+4b be the discriminant and let α and β be the roots of the characteristic polynomial x2axb. Then it is well known that if D0, then the Binet formula

    Un=αnβnαβ and Vn=αn+βn holds for all n0.

    Next, we recall Sanna's result [18] on the p-adic valuation of the Lucas sequence of the first kind.

    Lemma 3. [18,Theorem 1.5] Let p be a prime number such that pb. Then, for each positive integer n,

    vp(Un)={vp(n)+vp(Up)1if pD and p n,0if pD and p n,vp(n)+vp(Upτ(p))1if pD,τ(p)n,and pn,vp(Uτ(p))if pD,τ(p)n,and pn,0if pD and τ(p) n.

    In particular, if p is an odd prime such that pb, then, for each positive integer n,

    vp(Un)={vp(n)+vp(Up)1if p D and p  n,0if p D and p n,vp(n)+vp(Uτ(p))if p D andτ(p) n,0if p D and τ(p) n.

    From Lemma 3, and the fact that Vn=U2n/Un, we easily obtain the following result.

    Lemma 4. If p is an odd prime and pb. Then, for each positive integer n,

    vp(Vn)={vp(n)+vp(Uτ(p))if pD,τ(p)n and τ(p) 2n,0otherwise.

    Proof. This follows from the application of Lemma 3, a straightforward calculation, and the fact that vp(Vn)=vp(U2nUn)=vp(U2n)vp(Un).

    Next, we give some old and new lemmas that are needed in the proof of main theorems.

    Lemma 5. Let n1 and (a,b)=1. If pUn or pVn, then pb. Consequently, (Un,b)=(Vn,b)=1 for all n1.

    Proof. The case for Un is already given in [8,Lemma 7]. So suppose by way of contradiction that pVn and pb. Since Vn=aVn1+bVn2 and (a,b)=1, we obtain pVn1. Repeating this argument, we see that pVm for 1mn. In particular, pV1=a contradicting (a,b)=1. So if pVn, then pb, and the proof is complete.

    Lemma 6. [8,Lemma 8] Let a and b be odd, (a,b)=1, and v2(U6)v2(U3)+2. Then v2(U3)=1.

    For convenience, we also calculate the 2-adic valuation of Un and Vn as follows.

    Lemma 7. Assume that a is odd, b is even, and n1. Then v2(Un)=v2(Vn)=0.

    Proof. Since U1=1 and U2=a are odd, and Ur=aUr1+bUr2Ur1(mod2) for r3, it follows by induction that Un is odd. Since Vn=U2nUn, Vn is also odd. This proves this lemma.

    Lemma 8. Assume that a is even, b is odd, and n1. Then

    v2(Un)={v2(n)+v2(a)1if 2n,0if 2n,v2(Vn)={1if 2n,v2(a)if 2n,

    Proof. Since 2D, we obtain by Lemma 3 that for each nN, v2(Un)=v2(n)+v2(U2)1 if 2n and v2(Un)=0 if 2n. Since U2=a, the formula for v2(Un) is verified. Then v2(Vn) can be obtained from a straightforward calculation and the fact that Vn=U2nUn. This completes the proof.

    Lemma 9. Assume that a and b are odd, and n1. Then

    v2(Un)={v2(n)+v2(U6)1if n 0(mod6),v2(U3)if n 3(mod6),0if n 0(mod3),v2(Vn)={1if n 0(mod6),v2(U6)v2(U3)if n3(mod6),0if n0(mod3),

    Proof. Since U1 and U2 are odd, and U3=a2+b is even, we have τ(2)=3. In addition, 2D. Furthermore, 3n and 2n if and only if n0(mod6); 3n and 2n if and only if n3(mod6). Then applying Lemma 3 and the fact that Vn=U2nUn, we obtain the desired result.

    We begin with the simplest theorem of this paper.

    Theorem 10. Assume that k,m,nN, a,bZ, (a,b)=1, and m is odd. Then

    (i) if Vknm, then Vk+1nVnm;

    (ii) if Vknm, then Vk+1nVnm;

    (iii) if VknVnm, then Vk1nm;

    (iv) if VknVnm, then Vk1nm.

    Proof. We use Lemma 5 without reference. For (i), assume that Vknm. Since m is odd, Vn is also odd, and so v2(Vk+1n)=0. If p>2 and pVn, then pb and we obtain by Lemma 4 that

    vp(Vnm)=vp(mn)+vp(Uτ(p))=vp(m)+vp(n)+vp(Uτ(p))vp(Vkn)+vp(Vn)=vp(Vk+1n).

    Therefore vp(Vnm)vp(Vk+1n) for all primes p dividing Vn. This implies Vk+1nVnm.

    For (ii), assume that Vknm. By (i), it is enough to show that Vk+2nVnm. Since Vk+1nm, there exists a prime p dividing Vn such that vp(Vk+1n)>vp(m). Here we remark that the letter p in the proof of (i) and in the proof of (ii) may be different or may be the same. We believe that there is no ambiguity since (i) is already done. Now since Vknm and m is odd, Vn is also odd, and so v2(Vk+1n)=v2(m)=0. Therefore p is odd. By Lemma 4, we obtain

    vp(Vnm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(n)+vp(Uτ(p))=vp(m)+vp(Vn)<vp(Vk+1n)+vp(Vn)=vp(Vk+2n).

    This shows that Vk+2nVnm, as required.

    For (iii), assume that VknVnm. We show that vp(Vk1n)vp(m) for all primes p dividing Vn. If p is odd and pVn, then we apply Lemma 4 to obtain that

    vp(Vn)+vp(Vk1n)=vp(Vkn)vp(Vnm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(n)+vp(Uτ(p))=vp(m)+vp(Vn),

    and so vp(Vk1n)vp(m). It remains to show that v2(Vk1n)v2(m). If a is odd and b is even, then it follows from Lemma 7 that v2(Vk1n)=0v2(m). Recall that (a,b)=1, so a and b cannot be both even. So we have the following two remaining cases: (a is even and b is odd) or (a and b are odd).

    Case 1 a is even and b is odd. We will show that k must be 1, and so v2(Vk1n)=0v2(m). If 2n, then we apply Lemma 8 and the assumption that VknVnm to obtain

    1k=v2(Vkn)v2(Vnm)=1.

    Similarly, if 2n, then 2nm and we can use Lemma 8 again to obtain

    kv2(a)=v2(Vkn)v2(Vnm)=v2(a).

    In any case, k=1, as asserted.

    Case 2 a and b are odd. We use Lemma 9 in this case. If n0(mod3), then v2(Vk1n)=0v2(m). If n0(mod6), then nm0(mod6), and so k=v2(Vkn)v2(Vnm)=1; thus v2(Vk1n)=0v2(m). We now suppose n3(mod6). Since m is odd, nm3(mod6). Therefore

    k(v2(U6)v2(U3))=v2(Vkn)v2(Vnm)=v2(U6)v2(U3).

    So k=1 and thus v2(Vk1n)=0v2(m). Hence vp(Vk1n)vp(m) for all primes p dividing Vn, as desired. This proves (iii).

    For (iv), assume that VknVnm. By (iii), we have Vk1nm. If Vknm, then we obtain by (i) that Vk+1nVnm which contradicts VknVnm. Therefore Vk1nm. This completes the proof.

    Remark 11. Let k,m,nN, a,bZ, (a,b)=1, and m is even. Let p be an odd prime dividing Vn. By Lemma 4, we have pD, τ(p)n and τ(p)2n. Since m is even and τ(p)2n, we obtain τ(p)mn. By Lemma 4, we have pVnm, and so VnVnm. This shows that m in Theorem 10 cannot be even unless Vn=2r for some rN.

    Remark 12. The argument in Remark 11 works provided that there exists an odd prime p dividing Vn. The case Vn=2k for some kN{0} may occur but it is very rare. For example, when a=b=1, we know from the result of Bugeaud, Mignotte, and Siksek [2] that Vn is 1 or is a power of 2 if and only if n=0,1,3. Therefore we do not consider this rare case in our theorems.

    Lemma 13. Let k,m,nN, a,bZ, and (a,b)=1. Suppose m is odd and there exists an odd prime p dividing Vn. Then vp(Unm)=0 and VnUnm.

    Proof. By Lemma 4, we have pD, τ(p)n and τ(p)2n. Therefore τ(p) is even and v2(τ(p))=v2(n)+1. So τ(p)nm. By Lemma 3, vp(Unm)=0. Therefore VnUnm.

    Lemma 14. Let k,m,nN, a,bZ, and (a,b)=1. Suppose there exists an odd prime pUn. Then vp(Vnm)=0 and UnVnm.

    Proof. By Lemma 3, we have (i) vp(Un)=vp(n)+vp(Up)1 if pD and pn, and (ii) vp(Un)=vp(n)+vp(Uτ(p)) if pD and τ(p)n. For (i), we have vp(Un)>0 and pD, and therefore vp(Vnm)=0 and UnVnm. For (ii), we have τ(p)nm and so vp(Vnm)=0 and UnVnm.

    Remark 15. By Lemma 13 and a reason similar to that in Remark 12, we do not consider the case where m is odd in Theorems 16 to 20. In addition, by Lemma 14, we do not study the divisibility relation such as UknVnm.

    We now have the exact divisibility results for Un and Vn separately. In the next theorem, we consider them together. In other words, we investigate the relations of the type Vcnm implies VdnUnm; and VcnUnm implies Vdnm. We divide the results into 5 theorems according to the parities of a and b. From this point on, we apply Lemma 5 without reference.

    Theorem 16. Suppose that k,m,nN, a,bZ, (a,b)=1, a is odd, b is even, and m is even. Then

    (i) if Vknm, then Vk+1nUnm;

    (ii) if Vknm, then Vk+1nUnm;

    (iii) if Vk+1nUnm, then Vknm;

    (iv) if Vk+1nUnm, then Vknm.

    Proof. For (i), assume that Vknm. We show that vp(Vk+1n)vp(Unm) for all primes p dividing Vn. By Lemma 7, we have v2(Vn)=0. So let p be an odd prime dividing Vn. By Lemma 4, pD, τ(p)n, and τ(p)2n. Then τ(p)nm. By Lemmas 3 and 4, we obtain

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(n)+vp(Uτ(p))vp(Vkn)+vp(n)+vp(Uτ(p))=vp(Vkn)+vp(Vn)=vp(Vk+1n),as required.

    For (ii), assume that Vknm. By (i), it is enough to show that Vk+2nUnm. Since Vk+1nm, there exists a prime p such that vp(Vk+1n)>vp(m). By Lemma 7, v2(Vk+1n)=0, and so p2. Since pVn, we know that pD and τ(p)nm. Therefore we obtain by Lemmas 3 and 4 that

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(n)+vp(Uτ(p))=vp(m)+vp(Vn)<vp(Vk+1n)+vp(Vn)=vp(Vk+2n),as desired.

    For (iii), assume that Vk+1nUnm. By Lemma 7, v2(m)0=v2(Vkn). If p is odd and pVn, then we apply Lemmas 3 and 4 again to obtain

    vp(Vn)+vp(Vkn)=vp(Vk+1n)vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(n)+vp(Uτ(p))=vp(m)+vp(Vn).

    This shows that vp(Vkn)vp(m) for every prime p dividing Vn. So Vknm.

    For (iv), suppose Vk+1nUnm. By (iii), it is enough to show that Vk+1nm. If Vk+1nm, we apply (i) to obtain Vk+2nUnm contradicting Vk+1nUnm. Therefore the proof is complete.

    Theorem 17. Assume that k,m,nN, a,bZ, (a,b)=1, a is even, b is odd and m is even. Let

    t=min({v2(n)+v2(a)2}{ypk p is an odd prime factor of Vn})andyp=vp(m)vp(Vn)for each odd prime p dividing Vn.

    Then

    (i) if Vknm and 2n, then Vk+1nUnm;

    if Vknm and 2n, then Vk+1n2Unm;

    if Vknm, 2n, and v2(m)v2(Vkn)+1, then Vk+1nUnm;

    if Vknm, 2n, and Vk+1n2m, then t0, v2(m)k, and Vk+t+1nUnm;

    (ii) if Vknm, 2n and Vk+1n2m, then Vk+1nUnm;

    (iii) if Vknm, 2n and Vk+1n2m, then Vk+t+1nUnm;

    (iv) if Vknm, 2n and v2(m)=v2(Vkn), then VknUnm;

    (v) if Vknm, 2n and v2(m)v2(Vkn)+1, then Vk+1nUnm.

    Proof. For (i), assume that Vknm. If p is an odd prime and pVn, then pD, τ(p)nm, and we can apply Lemmas 3 and 4, to obtain

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(n)+vp(Uτ(p))vp(Vkn)+vp(Vn)=vp(Vk+1).

    From this point on, we sometimes use Lemmas 3 and 4 without reference. Next, we consider v2(Vk+1n) and v2(Unm). If 2n, then we apply Lemma 8 to obtain

    v2(Unm)=v2(nm)+v2(a)1=v2(m)+v2(n)+v2(a)1v2(Vkn)+v2(n)+v2(a)1v2(Vkn)+1=v2(Vkn)+v2(Vn)=v2(Vk+1n).

    This implies the first part of (i). Since m is even, 2nm. So if 2n, then we can still apply Lemma 8 to obtain

    v2(Unm)=v2(nm)+v2(a)1=v2(m)+v2(a)1v2(Vkn)+v2(a)1=v2(Vkn)+v2(Vn)1=v2(Vk+1n2). (3.1)

    This implies the second part of (i). For the third part of (i), we assume that 2n and v2(m)v2(Vkn)+1, and then we repeat the argument used in the second part to obtain

    v2(Unm)=v2(m)+v2(a)1v2(Vkn)+v2(a)=v2(Vk+1n).

    Therefore vp(Unm)vp(Vk+1n) for all primes p, which implies the desired result. Next, we prove the last part of (i). Assume that Vknm, 2n, and Vk+1n2m. Since a and n are even, v2(n)+v2(a)20. In addition, vp(m)vp(Vkn)=kvp(Vn), and so ypk. Therefore t0 and t+1v2(n)+v2(a)1. By Lemma 8, we have vp(Vn)=1, and therefore vp(m)k and

    v2(Unm)=v2(nm)+v2(a)1=v2(m)+v2(n)+v2(a)1k+t+1=v2(Vk+t+1n).

    If p is an odd prime and pVn, then

    vp(Unm)=vp(m)+vp(n)+vp(Uτ(p))=vp(m)+vp(Vn)ypvp(Vn)+vp(Vn)=(yp+1)vp(Vn)(k+t+1)vp(Vn)=vp(Vk+t+1n).

    Hence vp(Unm)vp(Vk+t+1n) for all primes p dividing Vn. Thus Vk+t+1nUnm, as desired.

    Next, we prove (ii). Assume that Vknm, 2n and Vk+1n2m. By (i), it is enough to show that Vk+2nUnm. By Lemma 8, we know that v2(Vn)=1. Then v2(m)v2(Vkn)=v2(Vk+1n2). Since Vk+1n2m, there exists an odd prime p dividing Vn such that vp(Vk+1n)>vp(m). Then pD, τ(p)nm, and

    vp(Vk+2n)=vp(Vk+1n)+vp(Vn)>vp(m)+vp(Vn)=vp(m)+vp(n)+vp(Uτ(p))=vp(Unm).

    This implies Vk+2nUnm.

    For (iii), assume that Vknm, 2n, and Vk+1n2m. By (i), we obtain t0, v2(m)k, and Vk+t+1nUnm. So it remains to show that Vk+t+2nUnm. We first observe that since Vk+1n2m, we obtain vp(Vk+1n)vp(m) for every odd prime p. If v2(m)k+1, then v2(m)v2(Vk+1n) which implies Vk+1nm contradicting the assumption Vknm. Therefore v2(m)=k. Next, we show that Vk+t+2nUnm. If t=ypk for some odd prime p dividing Vn, then we apply Lemmas 3 and 4 to obtain

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(Vn)=(vp(m)vp(Vn)+1)vp(Vn)<(yp+2)vp(Vn)=(k+t+2)vp(Vn)=vp(Vk+t+2n),

    and so Vk+t+2nUnm. If t=v2(n)+v2(a)2, then we obtain by Lemma 8 that

    v2(Unm)=v2(nm)+v2(a)1=v2(m)+v2(n)+v2(a)1=k+t+1<v2(Vk+t+2n),

    and so Vk+t+2nUnm. This proves (iii).

    Next, we prove (iv). Assume that Vknm, 2n and v2(m)=v2(Vkn). By (i), we have Vk+1n2Unm. To show that VknUnm, it suffices to prove that v2(Vkn)v2(Unm). Recall from (3.1) in the proof of the second part of (i) that

    v2(Unm)=v2(m)+v2(a)1=v2(Vkn)+v2(a)1v2(Vkn),

    and

    v2(Unm)=v2(m)+v2(a)1=v2(Vkn)+v2(Vn)1<v2(Vk+1n).

    So VknUnm and Vk+1nUnm. Thus VknUnm.

    For (v), assume that Vknm, 2n, and v2(m)v2(Vkn)+1. By (i), it suffices to show that Vk+2nUnm. Since Vk+1nm, there exists a prime p dividing Vn such that vp(Vk+1n)>vp(m). If p=2, then we obtain by Lemma 8 that

    v2(Unm)=v2(m)+v2(a)1<v2(Vk+1n)+v2(Vn)1<v2(Vk+2n),

    and so Vk+2nUnm. If p>2, then we obtain

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(Vn)<vp(Vk+1n)+vp(Vn)=vp(Vk+2n),

    which implies Vk+2nUnm. This completes the proof.

    From this point on, we apply Lemmas 3, 4, 5, and 8 without reference.

    Theorem 18. Suppose that k,m,nN, a,bZ, (a,b)=1, a is even, b is odd, and m is even. Then

    (i) for all odd primes p, if vp(Vk+1n)vp(Unm), then vp(Vkn)vp(m);

    (ii) if Vk+1nUnm and 2n, then Vmin(k,v2(m))nm;

    if Vk+1nUnm and 2n, then Vmin(k,v2(m))nm;

    (iii) if Vk+1nUnm and 2n, then Vknm;

    (iv) if Vk+1nUnm, 2n and Vk+2n2Unm, then Vknm;

    (v) if Vk+1nUnm, 2n, and Vk+2n2Unm, then Vk+1nm.

    Proof. For (i), assume that p is an odd prime and vp(Vk+1n)vp(Unm). If pVn, then

    vp(Vn)+vp(Vkn)=vp(Vk+1n)vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(n)+vp(Uτ(p))=vp(m)+vp(Vn),

    which implies (i). By (i), we only need to consider the 2-adic valuation in the proofs of (ii), (iii), (iv), and (v).

    For (ii), assume that Vk+1nUnm and 2n. For convenience, let c=min(k,v2(m)). If v2(m)k, then v2(Vkn)=kv2(m), and so Vknm. If v2(m)<k, then v2(Vv2(m)n)=v2(m) and vp(Vv2(m)n)vp(Vkn)vp(m) for all odd primes p, and therefore Vv2(m)nm. In any case, we obtain Vcnm. This proves the first part of (ii). Suppose further that Vk+1nUnm but Vc+1nm. Then

    v2(m)v2(Vc+1n)=min(k,v2(m))+1,

    which implies c=k. Then Vk+1n=Vc+1nm. By (i) of Theorem 17, we obtain Vk+2nUnm contradicting Vk+1nUnm. This completes the proof of (ii).

    For (iii), assume that Vk+1nUnm and 2n. Then

    v2(a)+v2(Vkn)=v2(Vk+1n)v2(Unm)=v2(nm)+v2(a)1=v2(m)+v2(a)1.

    Therefore v2(Vkn)<v2(m), and so Vknm.

    For (iv), assume that Vk+1nUnm, 2n, and Vk+2n2Unm. By (iii), Vknm. If Vk+1nm, then we obtain from (i) of Theorem 17 that Vk+2n2Unm, a contradiction. So Vknm.

    For (v), assume that Vk+1nUnm, 2n, and Vk+2n2Unm. Then

    v2(Vk+1n)+v2(a)1=v2(Vk+2n)1v2(Unm)=v2(nm)+v2(a)1=v2(m)+v2(a)1,

    and so v2(Vk+1n)v2(m). Therefore Vk+1nm. If Vk+2nm, we obtain from (i) of Theorem 17 that Vk+3n2Unm, which implies Vk+2nUnm contradicting Vk+1nUnm. Therefore Vk+1nm and the proof is complete.

    Theorem 19. Suppose that k,m,nN, a,bZ, (a,b)=1, a and b are odd, and m is even. Let c=v2(U6)1,

    t=min({v2(n)+c1}{ypk p is an odd prime factor of Vn}),s=min({c1}{ypk p is an odd prime factor of Vn}),andyp=vp(m)vp(Vn)for each odd prime p dividing Vn.

    Then

    (i) if Vknm, then Vk+1nUnm;

    (ii) if Vknm and n0(mod3), then Vk+1nUnm;

    (iii) if Vknm, n0(mod6) and Vk+1n2m, then Vk+1nUnm;

    (iv) if Vknm, n0(mod6), and Vk+1n2m, then t0 and Vk+t+1nUnm;

    if Vknm,n0(mod6) and vk+1n2m, then Vk+t+1nUnm;

    (v) if Vknm, n3(mod6), 2a2+3b and Vk+1n2m, then Vk+1nUnm;

    (vi) if Vknm, n3(mod6), 2a2+3b, and Vk+1n2m, then s0 and Vk+s+1nUnm;

    if Vknm, n3(mod6), 2a2+3b and Vk+1n2m, then Vk+s+1nUnm;

    (vii) if Vknm, n3(mod6), 4a2+3b and Vk+1n2cm, then Vk+1nUnm;

    (viii) if Vknm, n3(mod6), 4a2+3b and Vk+1n2cm, then Vk+2n2cUnm;

    if Vknm, n3(mod6), 4a2+3b and Vk+1n2cm, then Vk+2n2cUnm.

    Proof. As usual, to prove that VdnUnm, we show that vp(Vdn)vp(Unm) for all primes p dividing Vn. Similarly, if we would like to prove that VdnUnm, then we show that vp(Vdn)>vp(Unm) for some prime p. If p is odd, then we apply Lemmas 3 and 4; if p=2, then we use Lemma 9; and we will do this without further reference. For (i), assume that Vknm. If p is odd and pVn, then

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(n)+vp(Uτ(p))vp(Vkn)+vp(Vn)=vp(Vk+1).

    So it remains to show that v2(Unm)v2(Vk+1n). If n0(mod3), then v2(Vk+1n)=0v2(Unm). So suppose that n0(mod3). Then nm0(mod6) and so

    v2(Unm)=v2(nm)+v2(U6)1v2(Vkn)+v2(n)+v2(U6)1. (3.2)

    Since U3=a2+b is even and U6=a(a2+3b)U3, we know that v2(U3)1 and v2(U6)1. So if n0(mod6), then v2(n)1 and (3.2) implies that

    v2(Unm)v2(Vkn)+v2(U6)v2(Vkn)+v2(Vn)=v2(Vk+1n).

    If n3(mod6), then (3.2) implies

    v2(Unm)v2(Vkn)+v2(U6)1v2(Vkn)+v2(U6)v2(U3)=v2(Vk+1n).

    In any case, v2(Unm)v2(Vk+1n). This proves (i).

    For (ii), assume that Vknm and n0(mod3). By (i), it is enough to show that Vk+2nUnm. Since Vk+1nm, there exists a prime p dividing Vn such that vp(Vk+1n)>vp(m). Since v2(Vk+1n)=0, we see that p2. Then

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(n)+vp(Uτ(p))<vp(Vk+1n)+vp(Vn)=vp(Vk+2n), as desired.

    For (iii), assume that Vknm, n0(mod6), and Vk+1n2m. By (i), it is enough to show that Vk+2nUnm. Since Vk+1n2m and v2(Vk+1n2)=v2(Vkn)v2(m), we see that there exists an odd prime p dividing Vn such that vp(Vk+1n)>vp(m). Then

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(Vn)<vp(Vk+1n)+vp(Vn)=vp(Vk+2n).

    Therefore Vk+2nUnm, as required.

    For (iv), we first assume that Vknm, n0(mod6), and Vk+1n2m. Since v2(n)1 and v2(U6)v2(U3)1, it is not difficult to see that t0. If p is an odd prime dividing Vn, then

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(Vn)ypvp(Vn)+vp(Vn)=(yp+1)vp(Vn)(k+t+1)vp(Vn)=vp(Vk+t+1n).

    In addition,

    v2(Unm)=v2(nm)+v2(U6)1=v2(m)+v2(n)+v2(U6)1v2(Vkn)+t+1=k+t+1=v2(Vk+t+1n).

    Therefore Vk+t+1nUnm. This proves the first part of (iv). Next, assume further that Vknm. It is enough to show that Vk+t+2nUnm. Recall that yp=vp(m)vp(Vn), so vp(m)<(yp+1)vp(Vn). So if t=ypk for some odd prime p dividing Vn, then

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(Vn)<(yp+2)vp(Vn)=(k+t+2)vp(Vn)=vp(Vk+t+2n),

    which implies Vk+t+2nUnm. So suppose t=v2(n)+v2(U6)2. Since Vk+1n2m, we see that vp(m)vp(Vk+1n) for all odd primes p. If v2(m)k+1, then v2(m)v2(Vk+1n), which implies Vk+1nm contradicting the assumption Vknm. Therefore v2(m)k. Then

    v2(Unm)=v2(nm)+v2(U6)1=v2(m)+v2(n)+v2(U6)1k+t+1<v2(Vk+t+2n).

    Therefore, Vk+t+2nUnm as required.

    For (v), assume that Vknm, n3(mod6), 2a2+3b, and Vk+1n2m. By (i), it suffies to show that Vk+2nUnm. Since U6=a(a2+3b)U3 and 2a2+3b, we obtain v2(Vn)=v2(U6)v2(U3)=1. Since Vk+1n2m and v2(Vk+1n2)=v2(Vkn)v2(m), there exists an odd prime p dividing Vn such that vp(Vk+1n)>vp(m). Therefore

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(Vn)<vp(Vk+1n)+vp(Vn)=vp(Vk+2n), as desired.

    For (vi), assume that Vknm, n3(mod6), 2a2+3b, and Vk+1n2m. Since a2+3b and U3 are even, and U6=a(a2+3b)U3, we have v2(U6)20. Since Vknm, we have ypk for all odd primes p dividing Vn. Therefore s0. By the same argument as in the proof of (v), we obtain v2(Vn)=1. In addition, v2(m)v2(Vkn)=k and vp(Vk+1n)=vp(Vk+1n2)vp(m) for every odd prime p. If Vknm and v2(m)k+1=v2(Vk+1n), then Vk+1nm which is a contradiction. Therefore,

     if Vknm, then v2(m)=k. (3.3)

    We will apply (3.3) later. For now, we only need to apply v2(m)k. We obtain

    v2(Unm)=v2(nm)+v2(U6)1=v2(m)+v2(U6)1k+v2(U6)1k+s+1=v2(Vk+s+1n).

    If p>2 and pVn, then

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(Vn)(yp+1)vp(Vn)(k+s+1)vp(Vn)=vp(Vk+s+1n).

    This implies Vk+s+1nUnm. Next, assume further that Vknm. It remains to show that Vk+s+2nUnm. By the definition of yp, we know that (yp+1)vp(Vn)>vp(m). So if s=ypk for some odd prime p dividing Vn, then

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(Vn)<(yp+2)vp(Vn)=(k+s+2)vp(Vn)=vp(Vk+s+2n),

    which implies Vk+s+2nUnm. By (3.3), we know that v2(m)=k. So if s=v2(U6)2, then

    v2(Unm)=v2(nm)+v2(U6)1=v2(m)+v2(U6)1=k+s+1<v2(Vk+s+2n).

    So in any case, Vk+s+2nUnm, as required.

    For (vii), we let c=v2(U6)1 and assume that Vknm, n3(mod6), 4a2+3b, and Vk+1n2cm. By (i), it is enough to show that Vk+2nUnm. Since 4a2+3b and U6=a(a2+3b)U3, we have v2(U6)v2(U3)+2. By Lemma 6, we obtain v2(U3)=1, and so v2(Vn)=v2(U6)v2(U3)=v2(U6)1=c. Since Vk+1n2cm and

    v2(Vk+1n2c)=(k+1)v2(Vn)v2(Vn)=v2(Vkn)v2(m),

    there exists an odd prime p dividing Vn such that vp(Vk+1n)>vp(m). Then

    vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(Vn)<vp(Vk+1n)+vp(Vn)=Vp(Vk+2n).

    Therefore Vk+2nUnm.

    For (viii), assume that Vknm, n3(mod6), 4a2+3b, and Vk+1n2cm. Then for each odd prime p dividing Vn, we have

    vp(Vk+1n)=vp(Vk+1n2c)vp(m). (3.4)

    Since 4a2+3b and U6=a(a2+3b)U3, we obtain v2(U6)v2(U3)+2. By the same argument as in the proof of (vii), we obtain v2(Vn)=v2(U6)1=c. Since Vknm, we see that v2(m)v2(Vkn)=kv2(Vn). If Vknm and v2(m)(k+1)v2(Vn), then vp(m)vp(Vk+1n) for all primes p, and so Vk+1nm, a contradiction. Therefore

    v2(m)kv2(Vn), (3.5)

    and

     if Vknm, then kv2(Vn)v2(m)<(k+1)v2(Vn). (3.6)

    We will apply (3.6) later. For now (3.5) is good enough. We obtain

    v2(2cUnm)=v2(U6)1+v2(Unm)=v2(U6)1+v2(nm)+v2(U6)1=2(v2(U6)1)+v2(m)2(v2(U6)1)+kv2(Vn)=2(v2(U6)1)+k(v2(U6)1)=(k+2)(v2(U6)1)=v2(Vk+2n).

    If p>2 and pVn, then

    vp(2cUnm)=vp(Unm)=vp(nm)+vp(Uτ(p))=vp(m)+vp(Vn)vp(Vk+1n)+vp(Vn)=vp(Vk+2n),

    where the last inequality is obtained from (3.4). This implies that Vk+2n2cUnm. So the first part of (viii) is proved. Next, assume further that Vknm. To prove the second part, it now suffices to show that Vk+3n2cUnm. We have

    v2(2cUnm)=v2(U6)1+v2(Unm)=v2(U6)1+v2(nm)+v2(U6)1=2(v2(U6)1)+v2(m)<2(v2(U6)1)+(k+1)(v2(U6)1)=(k+3)(v2(U6)1)=v2(Vk+3n),

    where the inequality is obtained form (3.6) and the fact that v2(Vn)=v2(U6)1. This completes the proof.

    Theorem 20. Suppose that k,m,nN, a,bZ, (a,b)=1, a and b are odd and m is even. Then

    (i) for every odd prime p dividing Vn, if vp(Vk+1n)vp(Unm), then vp(Vkn)vp(m);

    (ii) if Vk+1nUnm and n0(mod3), then Vknm;

    if Vk+1nUnm and n0(mod3), then Vknm;

    (iii) if Vk+1nUnm, n0(mod6), and v2(m)k, then Vknm;

    if Vk+1nUnm, n0(mod6), and v2(m)k, then Vknm;

    if Vk+1nUnm, n0(mod6), and v2(m)<k, then Vv2(m)nm;

    (iv) if Vk+1nUnm, n3(mod6), 2a2+3b, and v2(m)k, then Vknm;

    if Vk+1nUnm, n3(mod6), 2a2+3b, and v2(m)k, then Vknm;

    if Vk+1nUnm, n3(mod6), 2a2+3b, and v2(m)<k, then Vv2(m)2m;

    (v) if Vk+1nUnm, n3(mod6), and 4a2+3b, then Vknm;

    if Vk+1nUnm, n3(mod6), and 4a2+3b, then Vknm.

    Proof. We apply Lemmas 3, 4, and 9 throughout the proof without reference. For (i), assume that p is an odd prime dividing Vn and vp(Vk+1n)vp(Unm). Then

    vp(Vn)+vp(Vkn)=vp(Vk+1n)vp(Unm)vp(nm)+vp(Uτ(p))=vp(m)+vp(Vn),

    which implies (i). Therefore we only need to consider the 2-adic valuation in the proof of (ii) to (v).

    For (ii), assume that Vk+1nUnm and n0(mod3). Since v2(Vkn)=0v2(m), we obtain by (i) that Vknm. Suppose futher that Vk+1nUnm. If Vk+1nm, then (i) of Theorem 19 implies Vk+2nUnm, which contradicts Vk+1nUnm, and so Vknm.

    For (iii), assume that Vk+1nUnm and n0(mod6).

    Case 1 v2(m)k. Then v2(Vkn)=kv2(m). So we obtain by (i) that Vknm. If Vk+1nUnm, then we obtain by (i) of Theorem 19 that Vk+1nm, and so Vknm. This proves (iii) in the case v2(m)k.

    Case 2 v2(m)<k. For convenience, let d=v2(m). Since v2(Vdn)=d=v2(m) and vp(Vdn)vp(Vkn)vp(m) for every odd prime p dividing Vn, we obtain Vdnm. If Vd+1nm, then d+1=v2(Vd+1n)v2(m)=d, a contradiction. So Vdnm.

    For (iv), assume that Vk+1nUnm, n3(mod6), and 2a2+3b. Since U6=a(a2+3b)U3 and 2a2+3b, we obtain v2(Vn)=v2(U6)v2(U3)=1.

    Case 1 v2(m)k. Then v2(Vkn)=kv2(m), and so we obtain by (i) that Vknm. If Vk+1nUnm, then we obtain by (i) of Theorem 19 that Vknm. This proves (iv) in the case v2(m)k.

    Case 2 v2(m)<k. For convenience, let d=v2(m). Then v2(Vdn)=d=v2(m) and vp(Vdn)vp(Vkn)vp(m). Therefore Vdnm. If Vd+1nm, then d+1=v2(Vd+1n)v2(m)=d, a contradiction. Therefore Vdnm.

    For (v), assume that Vk+1nUnm, n3(mod6), and 4a2+3b. Since U6=a(a2+3b)U3 and 4a2+3b, we obtain v2(U6)v2(U3)+2. By Lemma 6, we have v2(U3)=1. Then v2(Vn)=v2(U6)v2(U3)=v2(U6)1 and

    v2(Vkn)+v2(Vn)=v2(Vk+1n)v2(Unm)=v2(nm)+v2(U6)1=v2(m)+v2(Vn).

    So v2(Vkn)v2(m). By (i), we obtain Vknm. If Vk+1nUnm, then we obtain by (i) of Theorem 19 that Vk+1nm, and so Vknm. This completes the proof.

    We obtain exact divisibility theorems for the Lucas sequences of the first and second kinds, which complete a long investigation on this problem from 1970 to 2021.

    We are grateful to both referees for their kind words and suggestions which improve the quality of this paper. Kritkhajohn Onphaeng receives a scholarship from Development and Promotion for Science and Technology Talents Project (DPST). Prapanpong Pongsriiam's research project is funded jointly by Faculty of Science Silpakorn University and the National Research Council of Thailand (NRCT), Grant Number NRCT5-RSA63021-02.

    The authors declare that there is no conflict of interests regarding the publication of this article.



    [1] A. Benjamin, J. Rouse, When does FLm divide Fn? A combinatorial solution, Proceedings of the Eleventh International Conference on Fibonacci Numbers and Their Applications, 194, Congressus Numerantium, 2009, 53–58.
    [2] Y. Bugeaud, M. Mignotte, S. Siksek, Classical and modular approaches to exponential Diophantine equations I. Fibonacci and Lucas perfect powers, Ann. Math., 163 (2006), 969–1018. doi: 10.4007/annals.2006.163.969
    [3] P. Cubre, J. Rouse, Divisibility properties of the Fibonacci entry point, Proc. Amer. Math. Soc., 142 (2014), 3771–3785. doi: 10.1090/S0002-9939-2014-12269-6
    [4] V. E. Hoggatt Jr., M. Bicknell-Johnson, Divisibility by Fibonacci and Lucas squares, Fibonacci Quart., 15 (1977), 3–8.
    [5] Y. Matijasevich, Enumerable Sets are Diophantine, Proc. Academy Sci. USSR, 11 (1970), 354–358.
    [6] Y. Matijasevich, My collaboration with Julia Robison, Math. Intell., 14 (1992), 38–45. doi: 10.1007/BF03024472
    [7] Y. Matijasevich, Hilbert's Tenth Problem, MIT Press, 1996.
    [8] K. Onphaeng, P. Pongsriiam, Exact divisibility by powers of the integers in the Lucas sequence of the first kind, AIMS Math., 5 (2020), 6739–6748. doi: 10.3934/math.2020433
    [9] K. Onphaeng, P. Pongsriiam, Subsequences and divisibility by powers of the Fibonacci numbers, Fibonacci Quart., 52 (2014), 163–171.
    [10] K. Onphaeng, P. Pongsriiam, The converse of exact divisibility by powers of the Fibonacci and Lucas numbers, Fibonacci Quart., 56 (2018), 296–302.
    [11] C. Panraksa, A. Tangboonduangjit, p-adic valuation of Lucas iteration sequences, Fibonacci Quart., 56 (2018), 348–353.
    [12] A. Patra, G. K. Panda, T. Khemaratchatakumthorn, Exact divisibility by powers of the balancing and Lucas-balancing numbers, Fibonacci Quart., 59 (2021), 57–64.
    [13] P. Phunphayap, P. Pongsriiam, Explicit formulas for the p-adic valuations of Fibonomial coefficients, J. Integer Seq., 21 (2018), Article 18.3.1.
    [14] P. Phunphayap, P. Pongsriiam, Explicit formulas for the p-adic valuations of Fibonomial coefficients II, AIMS Math., 5 (2020), 5685–5699. doi: 10.3934/math.2020364
    [15] P. Pongsriiam, Exact divisibility by powers of the Fibonacci and Lucas numbers, J. Integer Seq., 17 (2014), Article 14.11.2.
    [16] P. Pongsriiam, Fibonacci and Lucas numbers associated with Brocard-Ramanujan equation, Commun. Korean Math. Soc., 32 (2017), 511–522.
    [17] M. K. Sahukar, G. K. Panda, Diophantine equations with balancing-like sequences associated to Brocard-Ramanujan-type problem, Glas Mat., 54 (2019), 255–270. doi: 10.3336/gm.54.2.01
    [18] C. Sanna, The p-adic valuation of Lucas sequences, Fibonacci Quart., 54 (2016), 118–124.
    [19] J. Seibert, P. Trojovský, On divisibility of a relation of the Fibonacci numbers, Int. J. Pure Appl. Math., 46 (2008), 443–448.
    [20] C. L. Stewart, On divisors of Lucas and Lehmer numbers, Acta Math., 211 (2013), 291–314. doi: 10.1007/s11511-013-0105-y
  • This article has been cited by:

    1. Phakhinkon Napp Phunphayap, Prapanpong Pongsriiam, Divisibility of Fibonomial coefficients in terms of their digital representations and applications, 2022, 7, 2473-6988, 5314, 10.3934/math.2022296
    2. Prapanpong Pongsriiam, Sums of divisors on arithmetic progressions, 2024, 88, 0031-5303, 443, 10.1007/s10998-023-00566-x
    3. Lejla Smajlović, Zenan Šabanac, Lamija Šćeta, On the Hurwitz-Type Zeta Function Associated to the Lucas Sequence, 2022, 60, 0015-0517, 355, 10.1080/00150517.2022.12427451
  • Reader Comments
  • © 2021 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(2592) PDF downloads(97) Cited by(3)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog