Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Let $m_1,\ldots, m_n$ be pairwise coprime natural numbers $\geq 1$. We consider the product $$G(m_1,\ldots,m_n) := \prod_{i=1}^{n} \mathbb{Z} / m_i \mathbb{Z}.$$ We define $M(n)$ as the set $n$-tuple of natural numbers $\geq 1$ with the property that the entrys are pairwise coprime. We define $l : M(n) \rightarrow \mathbb{N}_0$ by $$l(m_1,\ldots,m_n) = \max_{(x_1,\ldots,x_n) \in G(m_1,\ldots,m_n)} \min \{ k \in \mathbb{N}_0 | x_i+k \neq 0 \forall 1 \leq i \leq n \},$$ hence $l(m_1,\ldots,m_n)$ may be regarded as the maximal distance of the elements of $G$ from the elements which have no identity in their entrys. It is clear, that $l(m_1,\ldots,m_n)$ is always a natural number.

The question is now: Does there exist a real number $r$, such that $$r \cdot n \cdot \ln (n) \geq \sup_{(m_1,\ldots,m_n) \in M(n)} l(m_1,\ldots,m_n)$$ holds for all $n >> 0$ and is it possible to take $r=2$ ?

share|improve this question
2  
Could you please clarify the definition of $l(m_i,...)$, there seems to be a typo. –  Gjergji Zaimi Mar 21 '11 at 17:10
    
Hi, I edited my question. Hopefully, the typos is clear now. –  tobias Mar 21 '11 at 17:28
    
I am afraid not. What is the role of k in the max min? –  quid Mar 21 '11 at 17:32
    
What do you mean by the role of k ? It should be an absolute constant. –  tobias Mar 21 '11 at 17:36
    
In the displayed formula where a max min appears you say k in N_0 but this k appears nowhere; except later a different(?) k appears. –  quid Mar 21 '11 at 17:41
show 7 more comments

1 Answer

up vote 2 down vote accepted

We have $l(m_1,\cdots,m_n)$ is at least equal to the maximum length of a sequence of consecutive integers each of which is divisible by some $m_i$. Let's suppose for a moment that $m_i$ is the $i$th prime. In "On the problem of Jacobsthal", Iwaniec proves that this number is $\ll (n\log n)^2$, while your bound would be much stronger, and if you see the lower bound I mention in this equvalent question ($\sim\frac{n(\log n)^2 \log \log \log n}{(\log \log n)^2}$), it is actually false. In any case the question reduces to finding the best upper bound on Jacobsthal's function (In the comments of the question I linked to above there is an argument for why it suffices to reduce to the $m_i$ prime case).

share|improve this answer
    
Thank you very much Gjergji. But naturally the most important example of this is, as you already said, when $m_i$ is the i-th prime. If we would regard $l(m_1,...,m_n)$ for this case, could we give a good bound for $l(m_1,...,m_n)$ ? –  tobias Mar 21 '11 at 22:52
    
When $m_i$ are prime then $l(m_i)$ is equal to Jacobsthal's function. Except for Iwaniec's bound I mentioned here, and Erdos' bound in the other question, I am not aware of further improvement on lower or upper bounds. –  Gjergji Zaimi Mar 21 '11 at 22:58
    
I never heard of the Jacobsthal function before. Thas helps me quite a lot. Thank you, Gjergji. –  tobias Mar 22 '11 at 6:35
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.