Function expansion help

I want to write a function, $f(k,a,b)$, I made, in terms of combinations of the fractional part function, $$j\left\{\frac{c \ }{d}k\right\},$$ where $c,d,$ and $j$ are any integers.

The function is as follows: $f(k,a,b)=1$ if $k\equiv b$ mod a

and $f(k,a,b)=0$, if it is not

I need a general method for finding the expansion of this function in terms of the fractional part function for any given coprime integers $a,b$.

An example of one is $$f(k,6,1)= -\{k/6 \}+\{k/2\}+\{2k/3\}$$ Such that if, $k\equiv 1$ mod 6 , then $f(k,6,1)=1$, if not $f(k,6,1)=0$.

So I need a general method for finding the expansion of $f(k,a,b)$ in terms of the fractional part operator.

-

2 Answers

Let C be a (m-1)*(m-1) matrix of numbers $\large\ c_{ij}=\left\{\Large\frac{i\cdot j}{m}\right\}$
Let $D=C^{-1}$

Then
$\large f(k,m,1)=\sum_{i=1}^{m-1}d_{i1}\Large\{\frac{i}{m}\cdot\large k\}$
$\large f(k,m,r)=f(k-r+1,m,1)$

Example: $\ \ \ \ \ m=6$ $$C=\frac{1}{6}\cdot\begin{pmatrix} 1 & 2 & 3 & 4 & 5 \\ 2 & 4 & 0 & 2 & 4 \\ 3 & 0 & 3 & 0 & 3 \\ 4 & 2 & 0 & 4 & 2 \\ 5 & 4 & 3 & 2 & 1 \end{pmatrix}$$ $$D=C^{-1}=\begin{pmatrix} -1 & 0 & 1 & 1 & 0 \\ 0 & 1 & -1 & -1 & 1 \\ 1 & -1 & 0 & -1 & 1 \\ 1 & -1 & -1 & 1 & 0 \\ 0 & 1 & 1 & 0 & -1 \end{pmatrix}$$ $$\large f(k,6,1)=(-1)\cdot\{\frac{1}{6}\cdot k\}+1\cdot\{\frac{3}{6}\cdot k\}+1\cdot\{\frac{4}{6}\cdot k\}$$

-

Recall that the fractional part function $\{x\}$ is defined as the unique real number $r$ with $0\le r<1$ such that $x-r$ is an integer. Recall also that the floor function $\lfloor x\rfloor$ is the unique integer $m$ such that $m\le x<m+1$. Finally, recall that these two functions are related by the identity $x=\lfloor x\rfloor+\{x\}$.

If $a=0$, then your condition for $f(k)$ to be $1$ is $k=0n+b=b$, so we may write:

$$f(k)=\left\lfloor\frac{1}{1+|k-b|}\right\rfloor=\frac{1}{1+|k-b|}-\left\{\frac{1}{1+|k-b|}\right\}.$$

This works because for all $k$, $|k-b|\ge 0$, and $|k-b|=0$ exactly when $k=b$. Thus whenever $k\ne b$, $1+|k-b|>1$ and $0<\frac{1}{1+|k-b|}<1$, so $\left\lfloor\frac{1}{1+|k-b|}\right\rfloor=0$, and if $k=b$, then $1+|k-b|=1$ and $\frac{1}{1+|k-b|}=1$, so $\left\lfloor\frac{1}{1+|k-b|}\right\rfloor=1$.

Otherwise, the condiation for $f(k)$ to be $1$ is $k=an+b$, which is equivalent to $k-b=an$ or $\frac{k-b}{a}=n$. That is, $f(k)=1$ precisely when $\frac{k-b}{a}$ is an integer. Therefore we may write:

$$f(k)=1-\left(\left\{\frac{k-b}{a}\right\}+\left\{-\frac{k-b}{a}\right\}\right).$$

This works because for all $x$, if $\{x\}>0$, then $0\le 1-\{x\}<1$ and $-x-(1-\{x\})=-x+\{x\}-1=-(x-\{x\})-1=-\lfloor x\rfloor-1$, which is an integer, so we must have $\{-x\}=1-\{x\}$ and thus $\{x\}+\{-x\}=1$. On the other hand, if $\{x\}=0$, then $x$ is an integer, so $-x$ is also an integer and we must have $\{-x\}=0$, so $\{x\}+\{-x\}=0$. Therefore $\{x\}+\{-x\}$ is $0$ if $x$ is an integer and $1$ otherwise, so $1-\left(\{x\}+\{-x\}\right)$ is $1$ if $x$ is an integer and $0$ otherwise. We may then substitute $\frac{k-b}{a}$ for $x$ and get the desired function.

-