Abstract
In this paper, we give an upper bound for the solutions x, y, z of the equation in the title, of magnitude \(\left( \log \max \{a, b, c\}\right) ^{2 + \epsilon }\). This yields an improvement of earlier results of Hu and Le, where the bound is cubic in \(\log \max \{a, b, c\}\).
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Let a, b and c be coprime integers greater than one and let (x, y, z) be any positive integral solution of the equation \(a^x + b^y = c^z.\) In 2015, Hu and Le [2] proved that \(\max \{x, y, z\} \le \kappa (\log \max \{a, b, c\})^{3}\), where \(\kappa = 155{,}000\). Recently, the same authors improved the result as \(\kappa = 6500\), in [3]. In this paper, we prove the following theorem:
Theorem 1
For each \(0<\epsilon < 1\), there exists an effectively computable constant \(\kappa (\epsilon )> 0\) such that
Moreover, if l, m and n are positive integers such that one of \((a\bmod 2, m, n)\), \((b\bmod 2, l, n)\) and \((c \bmod 2, l, m)\) is (0, 1, 1), then all positive integral solutions (x, y, z) of the equation
satisfy the inequality with \(\kappa (\epsilon ) = \kappa (\epsilon , l, m, n)\), an effectively computable constant depends only on \(\epsilon \), l, m and n.
2 Proof of Theorem 1
Though there are better lower bounds in the literature than what Lemma 1 gives, it is sufficient for the present purposes.
Lemma 1
(Corollary B.1 of Shorey-Tijdeman [4]) Let \(\alpha _1\), \(\alpha _2\), ..., \(\alpha _n\) be non-zero rational numbers of heights not exceeding \(A_1\), \(A_2\), ..., \(A_n\), respectively. We assume \(A_{j} \ge 3\) for \(1 \le j \le n\). Put
Then, there exist computable absolute constants \(\kappa _0\) and \(\kappa _1\) such that the inequalities
have no solution in rational integers \(\beta _1\), \(\beta _2\), ..., \(\beta _n\) of absolute values not exceeding \(B \ge 2\).
For a prime number p and a non-zero integer \(\alpha \), \({{\mathrm{ord}}}_{p}(\alpha )\) is the largest non-negative integer l such that \(p^l \mid \alpha \).
Lemma 2
(Bugeaud [1]) Let \(\alpha _1\) and \(\alpha _2\) be non-zero multiplicatively independent integers and let \(\beta _1\) and \(\beta _2\) be positive integers. Assume that there exist a positive integer g and a real number E such that
and
Let \(A_1 >1\) and \(A_2 > 1\) be real numbers such that
for \(i = 1, 2\). Put
Consider
Then, we have
if p is odd or if \(p =2\) and \({{\mathrm{ord}}}_{2}(\alpha _2 - 1) \ge 2\).
Fix \(\min \{x, y, z\} \ge 4\). First, we shall assume that c is even. Then a and b are odd. Here, we work modulo 4. Since \(a^x + b^y \equiv 0 \pmod {4}\) and so it is impossible that both x and y are even, \(\wedge ' = a^x + b^y\) can be written as \(\wedge ' = \alpha _{1}^{\beta _1} - \alpha _{2}^{\beta _2}\), where \(\alpha _{1}\), \(\alpha _{2}\), \(\beta _1\) and \(\beta _2\) are integers such that \(\alpha _1 \equiv \alpha _2 \equiv 1 \pmod {4}\) and \((\alpha _{1}, \alpha _{2}, \beta _1, \beta _2)\) is one of \((a, -b,x, y)\), \((b, -a, y,x)\), \((a^2,-b,x/2,y)\) and \((b^2,-a,y/2,x)\). Furthermore, take \(p=2\), \(E = 2\), \(g = 1\), \(A_1= \max \{4, |\alpha _{1}|\}\) and \(A_2 = \max \{4,|\alpha _2|\}\) in Lemma 2. Since \(\gcd (\alpha _1, \alpha _2) = 1\), by Lemma 2, we have
where \(\kappa _{2}\) is a computable absolute constant. Also, it is clear that \(z \le {{\mathrm{ord}}}_{2}(c^{z})\). Therefore, we conclude that
This provides the bound for \(\max \{x, y, z\}\), since \(a^x, b^y \le c^z\).
Next, assume that c is odd. Then a or b is even. Without loss of generality, suppose that a is even. Let \(\wedge '= c^z - b^y \). Then, similarly as above, we obtain
Here, we distinguish two subcases. If \(c^z \le a^{2x}\), then the last two inequalities along with the inequality \(b^y \le c^z\) give the required bound. Otherwise, the given equation implies that
Take \(n = 2\), \(\alpha _1 = c\), \(\alpha _2 = b\), \(\beta _1 = - z\), \(\beta _2 = y\), \(A_1 = \max \{3, c\}\), \(A_2 = \max \{3, b\}\) and \(B = \max \{y, z\}\) in Lemma 1. Therefore, by the lemma, we write
where \(\kappa _{3}\) is a computable absolute constant. Combine this inequality with \(c^z \ge a^{2x}\) and \(b^y \le c^z\) to have the bound.
Next, fix \(\min \{x, y, z\} \le 3\). Then, \(a^x \le c^z\) and \(b^y \le c^z\) provide the required bound, if \(\min \{x, y, z\} = z\). Suppose that \(\min \{x, y, z\} = x\). Then, we get \(c^z - b^y \le a^3\) and it can be written as that \(0 < c^z b^{-y} - 1 \le a^3 b^{-y}\) if \(\max \{y, z\} = y\) or that \(0 < 1 - b^y c^{-z} \le a^3 c^{-z}\) if \(\max \{y, z\} = z\). Now, apply Lemma 1 by taking \(n = 2\), \(\alpha _1 = b\), \(\alpha _2 = c\), \(A_1 = \max \{3, b\}\), \(A_2 = \max \{3, c\}\) and \(B = \max \{y, z\}\). Then, we have the bound. Also, similarly as above, we can deal with the case \(\min \{x, y, z\} = y\).
For the proof of the second part of the theorem, in the above arguments replace \(c^z\) by \(nc^z\) if c is even, \(a^x\) by \(la^x\) if c is odd and a is even, and \(b^y\) by \(mb^y\) if c is odd and b is even, respectively.
Acknowlegements
I wish to remember my Ph.D. research supervisors Prof. R. Srikanth (SASTRA University) and Prof. R. Thangadurai (Harish-Chandra Research Institute) with gratitude. I express my sincere thanks to the referees of this paper for their valuable and constructive suggestions.
References
Bugeaud, Y.: Linear forms in p-adic logarithms and the diophantine equation \((x^n - 1)/(x-1) = y^q\). Math. Proc. Cambridge Philos. Soc. 127, 373–381 (1999)
Hu, Y., Le, M.: A note on ternary purely exponential diophantine equations. Acta Arith. 171, 173–182 (2015)
Hu, Y., Le, M.: An upper bound for the number of solutions of ternary purely exponential diophantine equations. J. Number Theory 183, 62–73 (2018)
Shorey, T.N., Tijdeman, R.: Exponential Diophantine Equations. Cambridge University Press, Cambridge (1986)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Subburam, S. On the diophantine equation \(la^x + mb^y = nc^z\). Res. number theory 4, 25 (2018). https://doi.org/10.1007/s40993-018-0118-x
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s40993-018-0118-x