Three realization problems about univariate polynomials
Vladimir Petrov Kostov · 2026
Read Three realization problems about univariate polynomials by Vladimir Petrov Kostov (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-10529.txt.
Introduction
Vladimir Petrov Kostov takes up three build asks for real polynomials. Each has one variable and no zero terms. Its terms have signs that form a pattern. A build ask goes the other way. Given a sign pattern, can a polynomial with it be built?
Kostov works on three such asks. The pattern is read from the top term down to the fixed term. Each ask adds its own binds on the terms. The paper seeks which patterns can occur and which can not. Such asks tie sign facts to root facts in a sharp way.
Catalog record
| CREATOR | Vladimir Petrov Kostov |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 34 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-10529
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Three realization problems about univariate polynomials Vladimir Petrov Kostov Abstract . We consider three realization problems about monic real univariate polynomials without vanishing coefficients. Such a polynomial P := ∑ j = 0 d b j x j P:=\sum_{j=0}^{d}b_{j}x^{j} defines the sign pattern σ ( P ) := ( sgn ( b d ) CLOSE \sigma(P):=({\rm sgn}(b_{d}) , … \ldots , OPEN sgn ( b 0 ) ) {\rm sgn}(b_{0})) . The numbers p d p_{d} and n d n_{d} of positive and negative roots of P P (counted with multiplicity) satisfy the Descartes’ rule of signs. Problem 1 asks for which couples C C of the form (sign pattern σ \sigma , pair ( p d , n d ) (p_{d},n_{d}) compatible with σ \sigma in the sense of Descartes’ rule of signs), there exist polynomials P P defining these couples. Problem 2 asks for which d d -tuples of pairs T := ( ( p d , n d ) CLOSE T:=((p_{d},n_{d}) , … \ldots , OPEN ( p 1 , n 1 ) ) (p_{1},n_{1})) , there exist polynomials P P such that P ( d − j ) P^{(d-j)} has p j p_{j} positive and n j n_{j} negative roots. A d d -tuple T T determines the sign pattern σ ( P ) \sigma(P) , but the inverse is false. We show by an example that 6 6 is the smallest value of d d for which there exist non-realizable tuples T T for which the corresponding couples C C are realizable. The third problem concerns polynomials with all roots real. We give a geometric interpretation of the three problems in the context of degree 4 4 polynomials. Key words: real polynomial in one variable; hyperbolic polynomial; sign pattern; Descartes’ rule of signs AMS classification: 26C10; 30C15 1. Introduction The present paper concerns three realization problems inspired by Descartes’ rule of signs about real univariate polynomials, see [ 2 ] , [ 4 ] , [ 5 ] , [ 6 ] , [ 7 ] , [ 8 ] , [ 10 ] , [ 14 ] or [ 16 ] . For a degree d monic polynomial (1.1) P:=x^{d}+\sum_{j=0}^{d-1}b_{j}x^{j}~,~~~\,b_{j}\in\mathbb{R}~, one knows that the number \ell_{+} of its positive roots counted with multiplicity is bounded by the number \tilde{c} of sign changes in the sequence of its coefficients the difference \tilde{c}-\ell_{+} being even. When counting sign changes, zero coefficients are ignored. If P has no vanishing coefficients, then for the number \ell_{-} of negative roots and the number of sign preservations d-\tilde{c} one obtains similar conditions by considering the polynomial P(-x) . Thus when b_{j}\neq 0 , j=0 , \ldots , d-1 , Descartes’ rule of signs implies the following restrictions: (1.2) \begin{array}[]{llc}\ell_{+}\leq\tilde{c}~,&\tilde{c}-\ell_{+}\in 2\mathbb{N}\cup 0~,&{\rm sgn}(b_{0})=(-1)^{\ell_{+}}~,\\ \\ \ell_{-}\leq d-\tilde{c}~,&d-\tilde{c}-\ell_{-}\in 2\mathbb{N}\cup 0~.\end{array} These conditions are only necessary. A. Albouy and Y. Fu have formulated sufficient conditions as well (see [ 1 ] and the references therein), about a more general class of functions of the form Y:=a_{0}x^{\alpha_{0}}+a_{1}x^{\alpha_{1}}+\cdots+a_{n}x^{\alpha_{n}} in which \alpha_{0}<\cdots<\alpha_{n} are arbitrary real exponents and a_{j} are non-zero real numbers. Namely, for a prescribed set E of positive numbers with cardinality \ell_{+} (repetitions are allowed), and for prescribed set of numbers \alpha_{j} as above and choice of the signs of the coefficients a_{j} with \tilde{c} sign changes, where \tilde{c}-\ell_{+}\in 2\mathbb{N}\cup 0 , there exists a function Y whose set of positive roots counted with multiplicity is the set E . The sufficient conditions of Albouy and Fu concern only the positive roots. However it would be natural to consider the positive and negative roots at the same time. (In what follows we speak about polynomials P without vanishing coefficients.) Definition 1 . (1) A sign pattern of length d+1 is a sequence of d+1 signs + and/or - . We say that the polynomial P defines the sign pattern \sigma(P)~:=~(~+~,~{\rm sgn}(b_{d-1})~,~\ldots~,~{\rm sgn}(b_{0})~)~. (2) Given a sign pattern \sigma with \tilde{c} sign changes (and hence d-\tilde{c} sign preservations) we define its Descartes’ pair as (\tilde{c},d-\tilde{c}) . A compatible pair (compatible with \sigma in the sense of Descartes’ rule of signs) is any pair (\tilde{c}-2u~,~d-\tilde{c}-2v)~,~~~\,u,~v\in\mathbb{N}\cup 0~,~~~\,\tilde{c}-2u\geq 0~,~~~\,d-\tilde{c}-2v\geq 0~. A couple (sign pattern, compatible pair) is called a compatible couple . (3) A compatible couple (\sigma_{0},(\ell_{+},\ell_{-})) is realizable if there exists a polynomial P with \sigma(P)=\sigma_{0} , with exactly \ell_{+} positive and \ell_{-} negative distinct roots. Example 1 . For the sign pattern \sigma_{\dagger}:=(+,-,-,-,+) , its Descartes’ pair is (2,2) and its compatible pairs are (2,2) , (2,0) , (0,2) and (0,0) . Remark 1 . Each compatible couple of the form (sign pattern, Descartes’ pair) is realizable, see [ 13 , Part (1) of Remarks 2.2.3] . The first realization problem which we consider reads: Problem 1 . For each given degree d , determine which compatible couples are realizable. Definition 2 . For a given degree d , the \mathbb{Z}_{2}\times\mathbb{Z}_{2} - action on the set of compatible couples is defined by the two commuting involutions i_{m}:P(x)\mapsto(-1)^{d}P(-x)~~~\,{\rm and}~~~\,i_{r}:P(x)\mapsto x^{d}P(1/x)/P(0)~. Remark 2 . The involution i_{m} exchanges the components of compatible pairs and changes the sign of every second coefficient of P . Up to a simultaneous change of the components of the sign pattern, the involution i_{r} reads sign patterns from the right and preserves compatible pairs. The factors (-1)^{d} and 1/P(0) are introduced in order to preserve the set of monic polynomials. Each orbit of the \mathbb{Z}_{2}\times\mathbb{Z}_{2} -action consists of 4 or 2 compatible couples. All compatible couples of a given orbit are simultaneously realizable or not. Orbits of length 2 occur when the compatible couple is invariant under one of the involutions i_{r} or i_{m}i_{r} ; one has i_{m}(\sigma)\neq\sigma for any sign pattern \sigma . Notation 1 . For a given degree d , we denote by \Sigma_{m_{1},m_{2},\ldots,m_{n}} , m_{1}+\cdots+m_{n}=d+1 , the sign pattern beginning by m_{1} signs + followed by m_{2} signs - followed by m_{3} signs + etc. For d=1 , 2 and 3 , all orbits are realizable, see [ 13 , p. 17] . For d=4 and d=5 , the only non-realizable orbits are the ones of the compatible couples (1.3) (\Sigma_{1,3,1},(0,2))~~~\,{\rm and}~~~\,(\Sigma_{1,4,1},(0,3)) respectively, see [ 9 ] and [ 1 ] . For d=6 , there are exactly 4 such orbits, see [ 1 ] : (1.4) (\Sigma_{1,5,1},(0,2))~,~(\Sigma_{1,5,1},(0,4))~,~(\Sigma_{4,1,2},(2,0))~~~\,{\rm and}~~~\,(\Sigma_{2,4,1},(0,4))~. Problem 1 is completely solved for d\leq 8 and partial results are obtained for higher degrees, see about Problem A in [ 13 ] . To formulate the next problem we have to study sequences of compatible pairs (SCPs) . For a degree d polynomial P , we denote by p_{j} and n_{j} the numbers of positive and negative roots of P^{(d-j)} respectively (counted with multiplicity). Hence in particular p_{d}=\ell_{+} and n_{d}=\ell_{-} . When all real roots of P and of all its non-constant derivatives are distinct, we say that P realizes the SCP S_{d}~:=~((p_{d},n_{d})~,~(p_{d-1},n_{d-1})~,~\ldots~,~(p_{1},n_{1}))~. Clearly (p_{1},n_{1})=(1,0) or (0,1) . Rolle’s theorem imposes the following restrictions on the numbers p_{j} and n_{j} : (1.5) p_{j}\leq p_{j-1}+1~,~~~\,n_{j}\leq n_{j-1}+1~~~\,{\rm and}~~~\,p_{j}+n_{j}\leq p_{j-1}+n_{j-1}+1~. Remark 3 . The SCP S_{d} determines the sign pattern \sigma(P) via the rule sgn (b_{j})=(-1)^{p_{d-j}} . However one and the same sign pattern beginning with +