A number theoretic identity - MathOverflow most recent 30 from http://mathoverflow.net2013-05-21T04:04:58Zhttp://mathoverflow.net/feeds/question/16972http://www.creativecommons.org/licenses/by-nc/2.5/rdfhttp://mathoverflow.net/questions/16972/a-number-theoretic-identityA number theoretic identityunknown (google)2010-03-03T14:49:32Z2010-03-03T15:00:38Z
<p>Let $n$ be a positive integer such that $2n+1$ is prime.
Elements of the factor group <code>$G = Z^*_{2n+1}/H$</code>,
where $H = \{-1,+1\}$, may be taken to be
$1,2,\ldots,n$. For every $x \in Z^*_{2n+1}$, let $x' \in \{1,2,\ldots,n\}$ denote
its image in $G$. For every $\lambda \in\{2,\ldots,n\}$, let </p>
<p>$S_\lambda = \sum_{a=1}^n {\rm abs} (a-(a\lambda)')$, </p>
<p>where ${\rm abs}(\cdot)$ denotes the usual absolute value. For example, when $n=6$,
and $\lambda=3$,
$S_3 = \sum_{a=1}^6 {\rm abs}(a-(3a)') = {\rm abs}(1-3) + {\rm abs}(2-6) + {\rm abs}(3-4) +
{\rm abs}(4-1) + {\rm abs}(5-2) + {\rm abs}(6-5) = 14$. </p>
<p>Is it true that $S_\lambda = n(n+1)/3$, for every $\lambda \geq 2$?</p>