Source author record

Tianxin Cai

Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.

ResearcherUnclaimed source record

Catalog footprint

Research overview

14works
1topics
4close collaborators

Published work

14 publications

preprint2026arXiv

Cyclotomic Congruences and Lucas Sequences

In this paper, we extend the pp-adic valuations originally obtained by Carmichael for the sequences obtained by applying Möbius inversion to Lucas sequences to pp-adic congruences, from which we immediately derive corresponding congruences for Lucas sequences. As a corollary, we also establish some constraints on the entry point behavior of primes in Lucas sequences, on the basis of which we conjecture the presence of a strong Chebyshev-like bias in real regular Lucas sequences.

preprint2016arXiv

A congruence involving harmonic sums modulo pαqβp^αq^β

In 2014, Wang and Cai established the following harmonic congruence for any odd prime pp and positive integer rr, \begin{equation*} Z(p^{r})\equiv-2p^{r-1}B_{p-3} ~(\bmod ~ p^{r}), \end{equation*} where Z(n)=∑i+j+k=ni,j,k∈Pn1ijk Z(n)=\sum\limits_{i+j+k=n\atop{i,j,k\in\mathcal{P}_{n}}}\frac{1}{ijk} and Pn\mathcal{P}_{n} denote the set of positive integers which are prime to nn. In this note, we obtain a congruence for distinct odd primes p, qp,~q and positive integers α, βα,~β, \begin{equation*} Z(p^αq^β)\equiv 2(2-q)(1-\frac{1}{q^{3}})p^{α-1}q^{β-1}B_{p-3}\pmod{p^α} \end{equation*} and the necessary and sufficient condition for \begin{equation*} Z(p^αq^β)\equiv 0\pmod{p^αq^β}. \end{equation*} Finally, we raise a conjecture that for n>1n>1 and odd prime power pα∣∣np^α||n, α≥1α\geq1, \begin{eqnarray} \nonumber Z(n)\equiv \prod\limits_{q|n\atop{q\neq p}}(1-\frac{2}{q})(1-\frac{1}{q^{3}})(-\frac{2n}{p})B_{p-3}\pmod{p^α}. \end{eqnarray}

preprint2016arXiv

On the Lucas Property of Linear Recurrent Sequences

We say that an arithmetical function S:N→ZS:\mathbb{N}\rightarrow\mathbb{Z} has Lucas property if for any prime pp, \begin{equation*} S(n)\equiv S(n_{0})S(n_{1})\ldots S(n_{r})\pmod p, \end{equation*} where n=∑i=0rnipin=\sum_{i=0}^{r}n_{i}p^{i}, with 0≤ni≤p−1,n,ni∈N0 \leq n_{i} \leq p-1,n,n_{i}\in\mathbb{N}. In this note, we discuss the Lucas property of Fibonacci sequences and Lucas numbers. Meanwhile, we find some other interesting results.

preprint2016arXiv

On the Sign Changes of a Weighted Divisor Problem

Let S(x;a1q1,a2q2)=∑′mn≤xcos⁡(2πma1q1)sin⁡(2πna2q2)S\big(x; \frac{a_1}{q_1}, \frac{a_2}{q_2}\big)=\mathop{{\sum}'}_{mn\leq x} \cos\big(2πm\frac{a_1}{q_1}\big)\sin\big(2πn\frac{a_2}{q_2}\big) with x≥q1q2,1≤ai≤qix\geq q_1q_2, 1\leq a_i\leq q_i, and (ai,qi)=1(a_i, q_i)=1 (i=1,2i=1, 2). We study the sign changes of S(x;a1q1,a2q2)S\big(x; \frac{a_1}{q_1}, \frac{a_2}{q_2}\big), and prove that for a sufficiently large constant CC, S(x;a1q1,a2q2)S\big(x; \frac{a_1}{q_1}, \frac{a_2}{q_2}\big) changes sign in the interval [T,T+CT][T,T+C\sqrt{T}] for any large TT. Meanwhile, we show that for a small constant c′c', there exist infinitely many subintervals of length c′Tlog⁡−7Tc'\sqrt{T}\log^{-7}T in [T,2T][T,2T] where ±S(t;a1q1,a2q2)>c5(q1q2)34t14\pm S\big(t; \frac{a_1}{q_1}, \frac{a_2}{q_2}\big)> c_5 (q_1q_2)^\frac{3}{4}t^\frac{1}{4} always holds.

preprint2015arXiv

A congruence involving alternating harmonic sums modulo pαqβp^αq^β

In 2014, Wang and Cai established the following harmonic congruence for any odd prime pp and positive integer rr, \begin{equation*} \sum\limits_{i+j+k=p^{r}\atop{i,j,k\in \mathcal{P}_{p}}}\frac{1}{ijk}\equiv-2p^{r-1}B_{p-3} ~(\bmod ~ p^{r}), \end{equation*} where Pn\mathcal{P}_{n} denote the set of positive integers which are prime to nn. In this note, we obtain the congruences for distinct odd primes p, qp,~q and positive integers α, βα,~β, \begin{equation*} \sum\limits_{i+j+k=p^αq^β\atop{i,j,k\in\mathcal{P}_{pq}\atop{i\equiv j\equiv k\equiv 1\pmod{2}}}}\frac{1}{ijk}\equiv\frac{7}{8}(2-q)(1-\frac{1}{q^{3}})p^{α-1}q^{β-1}B_{p-3}\pmod{p^α} \end{equation*} and \begin{equation*} \sum\limits_{i+j+k=p^αq^β\atop{i,j,k\in \mathcal{P}_{pq}}}\frac{(-1)^{i}}{ijk} \equiv \frac{1}{2}(q-2)(1-\frac{1}{q^{3}})p^{α-1}q^{β-1}B_{p-3}\pmod{p^α}. \end{equation*} Finally, we raise a conjecture that for n>1n>1 and odd prime power pα∣∣np^α||n, α≥1α\geq1, \begin{eqnarray} \nonumber \sum\limits_{i+j+k=n\atop{i,j,k\in\mathcal{P}_{n}}}\frac{(-1)^{i}}{ijk} \equiv \prod\limits_{q|n\atop{q\neq p}}(1-\frac{2}{q})(1-\frac{1}{q^{3}})\frac{n}{2p}B_{p-3}\pmod{p^α} \end{eqnarray} and \begin{eqnarray} \nonumber \sum\limits_{i+j+k=n\atop{i,j,k\in\mathcal{P}_{n}\atop{i\equiv j\equiv k\equiv 1\pmod{2}}}}\frac{1}{ijk} \equiv \prod\limits_{q|n\atop{q\neq p}}(1-\frac{2}{q})(1-\frac{1}{q^{3}})(-\frac{7n}{8p})B_{p-3}\pmod{p^α}. \end{eqnarray}

preprint2015arXiv

Super congruences involving alternating harmonic sums modulo prime powers

In 2014, Wang and Cai established the following harmonic congruence for any odd prime pp and positive integer rr, \begin{equation*} \sum\limits_{i+j+k=p^{r}\atop{i,j,k\in \mathcal{P}_{p}}}\frac{1}{ijk}\equiv-2p^{r-1}B_{p-3} (\bmod p^{r}), \end{equation*} where Pn\mathcal{P}_{n} denote the set of positive integers which are prime to nn. In this note, we establish a combinational congruence of alternating harmonic sums for any odd prime pp and positive integers rr, \begin{equation*} \sum\limits_{i+j+k=p^{r}\atop{i,j,k\in \mathcal{P}_{p}}}\frac{(-1)^{i}}{ijk} \equiv \frac{1}{2}p^{r-1}B_{p-3} (\bmod p^{r}). \end{equation*} For any odd prime p≥5p\geq 5 and positive integers rr, we have \begin{align} &4\sum\limits_{i_{1}+i_{2}+i_{3}+i_{4}=2p^{r}\atop{i_{1}, i_{2}, i_{3}, i_{4}\in \mathcal{P}_{p}}}\frac{(-1)^{i_{1}}}{i_{1}i_{2}i_{3}i_{4}}+3\sum\limits_{i_{1}+i_{2}+i_{3}+i_{4}=2p^{r}\atop{i_{1}, i_{2}, i_{3}, i_{4}\in \mathcal{P}_{p}}}\frac{(-1)^{i_{1}+i_{2}}}{i_{1}i_{2}i_{3}i_{4}} \nonumber\\&\equiv\begin{cases} \frac{216}{5}pB_{p-5}\pmod{p^{2}}, if r=1, \\ \frac{36}{5}p^{r}B_{p-5}\pmod{p^{r+1}}, if r>1. \\ \end{cases}\nonumber \end{align} For any odd prime p>5p> 5 and positive integers rr, we have \begin{align} &\sum\limits_{i_{1}+i_{2}+i_{3}+i_{4}+i_{5}=2p^{r}\atop{i_{1}, i_{2}, i_{3}, i_{4}, i_{5}\in \mathcal{P}_{p}}}\frac{(-1)^{i_{1}}}{i_{1}i_{2}i_{3}i_{4}i_{5}}+2\sum\limits_{i_{1}+i_{2}+i_{3}+i_{4}+i_{5}=2p^{r}\atop{i_{1}, i_{2}, i_{3}, i_{4}, i_{5}\in \mathcal{P}_{p}}}\frac{(-1)^{i_{1}+i_{2}}}{i_{1}i_{2}i_{3}i_{4}i_{5}} \nonumber\\&\equiv\begin{cases} 12B_{p-5}\pmod{p}, if r=1,\\ 6p^{r-1}B_{p-5}\pmod{p^{r}}, if r>1. \end{cases}\nonumber \end{align}

preprint2014arXiv

A New Generalization of Fermat's Last Theorem

In this paper, we consider some hybrid Diophantine equations of addition and multiplication. We first improve a result on new Hilbert-Waring problem. Then we consider the equation \begin{equation} \begin{cases} A+B=C ABC=D^n \end{cases} \end{equation} where $A,B,C,D,n \in\ZZ_{+}$ and n≥3n\geq3, which may be regarded as a generalization of Fermat's equation xn+yn=znx^n+y^n=z^n. When gcd⁡(A,B,C)=1\gcd(A,B,C)=1, (1)(1) is equivalent to Fermat's equation, which means it has no positive integer solutions. We discuss several cases for gcd⁡(A,B,C)=pk\gcd(A,B,C)=p^k where pp is an odd prime. In particular, for k=1k=1 we prove that (1)(1) has no nonzero integer solutions when n=3n=3 and we conjecture that it is also true for any prime n>3n>3. Finally, we consider equation (1)(1) in quadratic fields Q(t)\mathbb{Q}(\sqrt{t}) for n=3n=3.

preprint2014arXiv

Perfect numbers and Fibonacci primes

In this paper, we introduce the concept of FF-perfect number, which is a positive integer nn such that ∑d∣n,d<nd2=3n\sum_{d|n,d<n}d^2=3n. We prove that all the FF-perfect numbers are of the form n=F2k−1F2k+1n=F_{2k-1}F_{2k+1}, where both F2k−1F_{2k-1} and F2k+1F_{2k+1} are Fibonacci primes. Moreover, we obtain other interesting results and raise a new conjecture on perfect numbers.

preprint2014arXiv

Perfect Numbers and Fibonacci Primes (II)

In this paper, we study the diophantine equation σ2(n)−n2=An+B{{σ}_{2}}(n)-{{n}^{2}}=An+B. We prove that except for finitely many computable solutions, all the solutions to this equation with (A,B)=(L2m,F2m2−1)(A,B)=({{L}_{2m}},F_{2m}^{2}-1) are n=F2k+1F2k+2m+1n={{F}_{2k+1}}{{F}_{2k+2m+1}}, where both F2k+1{{F}_{2k+1}} and F2k+2m+1{{F}_{2k+2m+1}} are Fibonacci primes. Meanwhile, we show that the twin primes conjecture holds if and only if the equation σ2(n)−n2=2n+5{{σ}_{2}}(n)-{{n}^{2}}=2n+5 has infinitely many solutions.

preprint2013arXiv

A variety of Euler's conjecture

We consider a variety of Euler's conjecture, i.e., whether the Diophantine system {n=a1+a2+⋯+as−1,a1a2⋯as−1(a1+a2+⋯+as−1)=bs\begin{cases} n=a_{1}+a_{2}+\cdots+a_{s-1}, a_{1}a_{2}\cdots a_{s-1}(a_{1}+a_{2}+\cdots+a_{s-1})=b^{s} \end{cases} has solutions n,b,ai∈Z+,i=1,2,…,s−1,s≥3.n,b,a_i\in\mathbb{Z}^+,i=1,2,\ldots,s-1,s\geq 3. By using the theory of elliptic curves, we prove that it has no solutions n,b,ai∈Z+n,b,a_i\in\mathbb{Z}^+ for s=3s=3, but for s=4s=4 it has infinitely many solutions n,b,ai∈Z+n,b,a_i\in\mathbb{Z}^+ and for s≥5s\geq 5 there are infinitely many polynomial solutions n,b,ai∈Z[t1,t2,…,ts−3]n,b,a_i\in\mathbb{Z}[t_1,t_2,\ldots,t_{s-3}] with positive value satisfying this Diophantine system.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.