arXiv Science⌕ Search

arXiv subjects

Musbahu Idris

Publications and source records attributed to Musbahu Idris.

3 recordsLinked to original sources

On Borwein polynomials with multiple roots outside the unit disk

A Borwein polynomial has coefficients in ${-1,0,1}$ and nonzero constant and leading coefficients. We seek Borwein polynomials having a multiple root outside the unit disk. To this end, we consider the powers $L^2$ and $L^3$ of Lehmer's polynomial $L$, which has a real root $τ>1$. Since $L$ has exceptionally small Mahler measure, the theorem of Bombieri and Vaaler guarantees the existence of Borwein multiples of these powers. Such multiples have $τ$ as a root of multiplicity at least 2 and 3, respectively. Mossinghoff (2003, Algorithm 3, p. 323) constructed examples that, under the substitution $x\mapsto -x$, yield such multiples. The main contribution of this paper is to determine the least possible degrees of Borwein multiples of $L^2$ and $L^3$, with and without symmetry restrictions on their coefficients, and to describe their degree spectra. For $L^2$, the minimum degree is 34 without symmetry restrictions, 38 among self-reciprocal multiples, and 41 among antireciprocal multiples. For $L^3$, the corresponding minimum degrees are 115, 115, and 117. We determine the complete degree spectra for $L^2$ and, for $L^3$, classify the degrees from 115 to 150 for which self-reciprocal or antireciprocal Borwein multiples exist, giving an explicit example for each admissible degree in each symmetry class. Further constructions yield Borwein multiples of $L^3$ in every degree $D\geq 215$, with both symmetry types for every $D\geq 231$.

math.NT↗

On the existence of Newman and Littlewood multiples for certain integer polynomials

Newman polynomials have coefficients in {0, 1} and constant term 1, whereas Littlewood polynomials have coefficients in {-1, 1}. We study two questions concerning the existence of Newman and Littlewood multiples for certain integer polynomials. For Newman multiples, we revisit a question of Hare and Mossinghoff [6]: whether there exists a real number sigma > 1 such that every P in Z[x] with no nonnegative real root and Mahler measure less than sigma has a Newman multiple. To bound any possible value of sigma from above, we seek polynomials with no nonnegative real root and no Newman multiple, and with Mahler measure as small as possible. Using the updated database of known small-Mahler-measure polynomials up to degree 200 [11], we test, for each listed polynomial p(x), both p(x) and its sign transform p(-x). These two polynomials have the same Mahler measure, but the existence of a Newman multiple is not invariant under the substitution x -> x. After a degree-bounded prefilter formulated as a mixed-integer linear optimization problem, we apply the Hare-Mossinghoff certification procedure to the remaining candidates. This yields 14 irreducible polynomials of Mahler measure less than 1.3, with no nonnegative real root and no Newman multiple. The smallest of their Mahler measures is 1.263095875491..., showing that any such sigma is at most this value. For Littlewood multiples, we return to the three Newman polynomials listed in Table 3 of Drungilas, Jankauskas, Junevicius, Klebonas and Siurys [3], for which the existence of a Littlewood multiple of smallest possible degree was left unresolved. We show that one of them attains this degree, whereas the other two have no Littlewood multiple at either of their first two possible degrees.

math.NT↗

Algorithmic aspects of Newman polynomials and their divisors

We study the problem of determining which integer polynomials divide Newman polynomials. In this vein, we first give results concerning the $8438$ known polynomials with Mahler measure less than $1.3$. We then exhibit a list of polynomials that divide no Newman polynomial. In particular, we show that a degree-10 polynomial of Mahler measure \text{approximately} 1.419404632 divides no Newman polynomial, thereby improving the best known upper bound for any universal constant $σ$, if it exists, such that every integer polynomial of Mahler measure less than $σ$ divides a Newman polynomial. Finally, letting $l(x)$ denote Lehmer's polynomial, we explicitly construct Newman polynomials divisible by $l(x)^2$ with degrees up to $150$, and show that no Newman polynomial is divisible by $l(x)^3$ up to degree $160$.

math.NT↗