A minimum problem for finite sets of real numbers with nonnegative sum

are negative. Inspired by some interesting extremal combinatorial sum problems raised by Manickam, Mikl"os and Singhi in 1987 cite{ManMik87} and 1988 cite{ManSin88} we study the following two problems:

oindent $(P 1)$ {it which are the values of $g a m m a (n, r)$ and $e t a (n, r)$ for each $n$ and $r$ , $0 < r l e n$ ?} oindent $(P 2)$ {it if $q$ is an integer such that $g a m m a (n, r) l e q l e e t a (n, r)$ , can we find $n$ real numbers $a_{1}, c d o t s, a_{n}$ , such that $r$ of them are non-negative and the remaining $n - r$ are negative with $s u m_{1 = 1}^{n} a_{i} g e 0$ , such that the number of the non-negative sums formed from these numbers is exactly $q$ ?}

oindent We prove that the solution of the problem

(P 1)

is given by

g a m m a (n, r) = 2^{n - 1}

and

e t a (n, r) = 2^{n} - 2^{n - r}

. We provide a partial result of the latter problem showing that the answer is affirmative for the weighted boolean maps. With respect to the problem

(P 2)

such maps (that we will introduce in the present paper) can be considered a generalization of the multisets

a_{1}, c d o t s, a_{n}

with

s u m_{1 = 1}^{n} a_{i} g e 0

. More precisely we prove that for each

q

such that

g a m m a (n, r) l e q l e e t a (n, r)

there exists a weighted boolean map having exactly

q

positive boolean values.

Recommendations

Cites work

Cited in

(9)

This page was built for publication: A minimum problem for finite sets of real numbers with nonnegative sum

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q443111)