# Counting instances where $bn_k$ is equal to any $an_k-1$ or $an_k+1$ under a given number on a number-line.

If I let $a, b$, and $n$ be integers greater than $0$ and I incrementally and consecutively plot $a{n}-1$ and $a{n}+1$ on a number-line such that when $a=2$ , I plot points { $2n_1-1, 2n_1+1$, $2n_2-1, 2n_2+1$, ... , $2n_k-1,2n_k+1$ } and when $a=3$, I plot { $3n_1-1, 3n_1+1$, $3n_2-1, 3n_2+1$, ... , $3n_k-1, 3n_k+1$ }, possibly omitting duplicate points. Then I plot $bn$ along the same number-line as in { $bn_1, bn_2,...,bn_k$}. How do I count all the instances where any $bn_k$ is also equal to any $an-1$ or $an+1$ under a given number? Lastly, please excuse any informal use of notation resulting from my lack of knowledge in the topic and feel free to edit.

Clarification:You can ignore the confusing upper part of this question. I will be keeping it until the one answer I've gotten so far gets edited.

Consider the following graphic:

The points on the upper line are $12$ apart while the ones on the lower line are $5$ apart. The points on the upper line progress by $12n$ where each point's $n$ value is 1 more then the $n$ value of the previous point as in $\{12\ast1, 12\ast2, 12\ast3, ..., 12\ast k\}$. This mechanism applies to the lower line where the constant is $5$. Now, notice the two read points on the lower line which are $5\ast5$ and $5*7$, these points can be expressed as $12n+1$ and $12n+1$. My question is, given any length of plot how do i calculate the number of points that can be expressed either as $xn-1$ or $xn+1$. The two red points on the $5$ line can be expressed as $12n-1$ and $12n+1$ and the two black points on the twelve line next to the red points can be expressed as $5n-1$ and $5n+1$.

-
You haven't told us what $n_1,n_2,\dots,n_k$ are, nor what $k$ is. –  Gerry Myerson Jan 17 '13 at 22:41
Sorry about that; I've edited the question accordingly. –  Babiker Jan 17 '13 at 23:15
Still not clear. When you write $n=7$, do you mean $a=7$? More importantly, is my answer relevant, or have I missed the point? –  Gerry Myerson Jan 17 '13 at 23:42
You're right, I meant $a=7$. And yes, your answer is relevant. –  Babiker Jan 17 '13 at 23:51

The question is not clear.

If $a$ and $b$ are not relatively prime, you will never have $bn_k=an\pm1$. I don't understand the phrase, "under a given number".

So, let's assume $a$ and $b$ are relatively prime.

Then the equation $ax-by=1$ has exactly one solution in integers $x$ and $y$ with $1\le x\le b$ and $1\le y\le a$. More generally, it has exactly $Q$ solutions with $1\le x\le Qb$ and $1\le y\le Qa$. Note that $ax-by=1$ is $by=ax-1$ which is just $bn_k=an-1$ with different letters.

Similar formulas apply to $bn_k=an+1$, treating it as $bx-ay=1$.

If I have seriously misunderstood the question, please edit the question to clarify.

-
Hello there, can you help me out here? I constructed a complete model of the markov chain taxi cab problem with the two zones and then the question got deleted before I could post. What to do? –  Marko Riedel Jan 18 '13 at 0:51
@Marko, what do you want to do? OP got what was needed, and deleted the post. You could send email to the moderator team, I suppose, and tell them it was question 280938. Or you could start a thread on meta, explaining why you'd like people to vote to undelete. –  Gerry Myerson Jan 18 '13 at 2:20
Well I don't know if its worth all that trouble, it is too bad that we cannot put a temporary advisory lock on a question so one gets a warning when it is about to be deleted and can then take it to the comments section. I can however say in all honesty that I do not know whether my answer, which uses generating functions is (a) correct and cannot (b) be done more easily with Markov chains. So I could write an intro and re-post it as a question. However I do have a fear of being flamed as a spammer, so please advise me whether this would be acceptable. Then upvote/downvote after that. –  Marko Riedel Jan 18 '13 at 2:31
@Marko, genuine mathematical questions are never considered spam. If you have a way to do a problem, and you want to know whether it is correct and/or whether it's as good as some other way to do the problem, I can't think of a better way to find out than to post it as a new question at this website. –  Gerry Myerson Jan 18 '13 at 2:47
That's very kind. I think I'll give it another try. –  Marko Riedel Jan 18 '13 at 2:54