On a theorem of W.A. Markov

Abstract

Authors

T. Popoviciu
Institutul de Calcul

Keywords

?

Paper coordinates

T. Popoviciu, Asupra unei teoreme a lui W.A. Markov, Acad. R. P. Romîne, Fil. Cluj, Stud. Cerc. Mat., 12 (1961), pp. 333-355 (in Romanian).

PDF

About this paper

Journal

Studii si Cercetari Matematice

Publisher Name

Academy of the Republic of S.R.

Print ISSN

1220-269X

Online ISSN

google scholar link

??

Paper (preprint) in HTML form

Original text
Rate this translation
Your feedback will be used to help improve Google Translate

ON A THEOREM OF WA MARKOV

by
TIBERIU POPOVICIU

MATHEMATICA

VOLUME 2 (25)
FASCICOL 2

1960
LUCUJ{}^{\mathrm{C}}\mathrm{L}^{\mathrm{U}}\mathrm{U}\mathrm{J}

SOCIETA'TEA DE ŞTIIN'PE MATEMATICE ŞI FIZICE DIN RPR FILIALAA CI,UJ

MATHEMATICA

Vol. 2 (25), fasc. 2
1960

In Cluj,
W. A. ​​Markov, in his work [2] on the generalization of the famous AA Markov inequality, gave, as a preliminary lemma, the following theorem:

If the roots of two polynomials of degreenn, having all their real roots, separate, the same is true for the roots of the derivatives of these polynomials.

In the second part of this work we will give a proof of this theorem. Our proof differs slightly from that of Wa Markov and also from that of P. Monteil [3] given almost 30 years ago in this same journal.

The demonstration we present is based on the continuity and monotonicity of the roots of the derivative of a polynomial with all its roots being real, with respect to the roots of the polynomial. In the first part of this work, we will analyze this property of monotonicity.

Finally, in the third part of this work, we will give a new theorem on polynomials having all their roots real, analogous to that of WA Markov, cited above.

We consider only polynomials of one variable whose roots are all real, and by the degree of a polynomial we always mean its actual degree, even if these properties are not explicitly stated. Accents denote derivatives. We can consider as equal two polynomials that differ only by a non-zero multiplicative constant.

  1. 1.

    If a polynomial has all its roots real, its derivative also has all its roots real. There is an important, well-known property that separates the roots of the derivative from those of the polynomial. We will use this property in what follows.

The roots of a polynomial whose highest coefficient is equal to 1 (i.e., of the formxn+x^{n}+degree polynomial<n<nThe coefficients of the polynomial are continuous functions with respect to the coefficients, and the coefficients are continuous functions (of the polynomials) with respect to the roots of the polynomial. Considering the relationships between the roots and coefficients of a polynomial, we deduce the continuity of the roots of the derivative with respect to the roots of the polynomial.
2. The monotonicity property of the roots of the derivative with respect to the roots of the polynomial can be stated as follows:

The roots of the derivative are non-decreasing functions of the roots of the polynomial.

This property is well known and has been used a lot, for example by Laguerre in his research on polynomials having all their roots real.

To better highlight the monotonicity property, we introduce the relationshipPcQP\xrightarrow{c}Qbetween two polynomials, which occurs if and only if:
101^{0}I polynomialsP,QP,Qare of the same degreen1n\geqq 1.
22^{\circ}The respective roots

x1x2xn,y1y2ynx_{1}\leqq x_{2}\leqq\ldots\leq x_{n},y_{1}\leq y_{2}\leq\ldots\leq y_{n} (1)

of these polynomials satisfy the inequalities

xiyi,i=1,2,,n.x_{i}\leqq y_{i},i=1,2,\ldots,n. (2)

This relationship is (reflexive and) transitive. It is unnecessary to consider the casen=0n=0when the previous relationship makes no sense. Ifn=1n=1It suffices to mention only the inequality (2) of definition and we see that in this case at least one of the relationsPcQ,QcPP\stackrel{{\scriptstyle\mathrm{c}}}Q,Q\stackrel{{\scriptstyle\mathrm{c}}}Pis still true. For alln>1n>1We can construct polynomialsP,QP,Qsuch that none of the relationshipsPcQ,QcPP\xrightarrow{\mathrm{c}}Q,Q\xrightarrow{\mathrm{c}}Pthat it is not true.
I1hasI_{1}aThe monotonicity property of the roots of the derivative with respect to those of the polynomial is then expressed by the

THEOREM 1. IfPP,QQare two polynomials of degreen>1n>1, ofPcQP\xrightarrow{\mathrm{c}}Qit follows thatPcQP^{\prime}\xrightarrow{\mathrm{c}}Q^{\prime}.

We also introduce the relationshipPccQP\xrightarrow{cc}Qbetween two polynomials, which occurs if and only if:
11^{\circ}PolynomialsP,QP,Qare of the same degreen1n\geqq 1and both have very simple roots.
22^{\circ}The respective roots
(1')

x1<x2<<xn,y1<y2<<ynx_{1}<x_{2}<\ldots<x_{n},y_{1}<y_{2}<\ldots<y_{n}

of these polynomials satisfy the inequalities

xi<yi,i=1,2,,n.x_{i}<y_{i},\quad i=1,2,\ldots,n. (\prime)

This relation, which is also transitive, is a special case of the previous relation and is obtained when everywhere in (1) and (2) the sign\lesssimis replaced by<<.

The two relations considered are also linked by a kind of mixed transitivity, analogous to the corresponding property of inequality relations<<And\leqq. IfPccR,RcQP\xrightarrow{\mathrm{cc}}R,R\xrightarrow{\mathrm{c}}Qand ifQQto all its simple roots, we havePccQP\xrightarrow{\mathrm{cc}}Q. OfPccR,RcS,SccQP\xrightarrow{\mathrm{cc}}R,R\xrightarrow{\mathrm{c}}S,S\xrightarrow{\mathrm{cc}}Qit follows thatPccQP\xrightarrow{\mathrm{cc}}Q.

If the roots ofP,QP,Qare continuous functions of a parameterλ\lambdaon an interval that contains the pointλ0\lambda_{0}and if the relationshipPccQP\xrightarrow{cc}Qis checked forλλ0\lambda\neq\lambda_{0}, We havePcQP\stackrel{{\scriptstyle c}}Qbut not in generalPccQP\xrightarrow{cc}Q, Forλ=λ0\lambda=\lambda_{0}This property also holds true for pairs of relationships.s,ss;m;mn,mc\xrightarrow[\rightarrow]{s},\xrightarrow{ss};\xrightarrow{m};\xrightarrow{mn},\xrightarrow{mc}which we will introduce later.

We have 1st
THLOREME 2. IfPP,QQare two polynomials of degreen>1n>1, ofPQP\rightarrow Qit follows thatPcQP^{\prime}\xrightarrow{\mathrm{c}}Q^{\prime}3.
We will first demonstrate that Theorem 1 follows from Theorem 2. Indeed, letPcQP\xrightarrow{c}Q(1) the roots of the polynomialsP,Q,n>1P,Q,n>1And

ξ1ξ2ξn1,η˙1η2ηn1\xi_{1}\leqq\xi_{2}\leqq\cdots\leqq\xi_{n-1},\dot{\eta}_{1}\leqq\eta_{2}\leqq\cdots\leqq\eta_{n-1} (3)

the respective roots of the polynomialsP,QP^{\prime},Q^{\prime}Let us consider polynomialsPε,QεP_{\varepsilon},Q_{\varepsilon}degreenn, having respectively the rootsxi+iε,i=1,2,nx_{i}+i\varepsilon,i=1,2,\ldots n,yi+(i+1)ε,i=1,2,,ny_{i}+(i+1)\varepsilon,i=1,2,\ldots,nOrε\varepsilonis a positive number. Polynomials
Pε,QεP_{\varepsilon},Q_{\varepsilon}all have their simple roots, we havePεccQeP_{\varepsilon}\xrightarrow{\mathrm{cc}}Q_{e}and if

ξ1(e)<ξ2(e)<<ξn1(e),η1(e)<η2(e)<<ηn1(e)\xi_{1}^{(\mathrm{e})}<\xi_{2}^{(\mathrm{e})}<\ldots<\xi_{n-1}^{(\mathrm{e})},\quad\eta_{1}^{(\mathrm{e})}<\eta_{2}^{(\mathrm{e})}<\ldots<\eta_{n-1}^{(\mathrm{e})}

are respectively the roots of the polynomialsPe,QeP_{e}^{\prime},Q_{e}^{\prime}, We have

limitε0ξi(ε)=ξi,limitε0ηi(ε)=ηi,i=1,2,,n1.\lim_{\varepsilon\rightarrow 0}\xi_{i}^{(\varepsilon)}=\xi_{i},\lim_{\varepsilon\rightarrow 0}\eta_{i}^{(\varepsilon)}=\eta_{i},\quad i=1,2,\ldots,n-1. (4)

If we assume that Theorem 2 is true, it follows thatPεccQεP_{\varepsilon}^{\prime}\xrightarrow{cc}Q_{\varepsilon}^{\prime}and from (4) we deduce, byε0,PcQ\varepsilon\rightarrow 0,P^{\prime}\xrightarrow{\mathrm{c}}Q^{\prime}We have thus demonstrated that Theorem 1 is a continuation of Theorem 2.
4. It remains to prove Theorem 2. LetP,QP,Qtwo polynomials of degreen>1n>1such asPccQP\xrightarrow{cc}Qand be(1)\left(1^{\prime}\right)the respective roots of these polynomials. LetPiP_{i}a polynomial of degreennhaving as its rootsx1,x2,,xni,yni+1,yni+2,,ynx_{1},x_{2},\ldots,x_{n-i},y_{n-i+1},y_{n-i+2},\ldots,y_{n}, Fori=1,2,,ni=1,2,\ldots,nThe polynomialP0P_{0}is equal toPPAndPnP_{n}is equal toQQYour polynomialsPiP_{i}all have their simple roots, but we generally only havePi+cPi+1,i=0,1P_{i\rightarrow+}^{\mathrm{c}}P_{i+1},i=0,1,.n1\ldots.n-1If we demonstrate that

PiccPi+1,i=0,1,,n1,P_{i}^{\prime}\xrightarrow{\mathrm{cc}}P_{i+1}^{\prime},\quad i=0,1,\ldots,n-1, (5)

Therefore, due to the transitivity of the relation in question, it follows thatPccQP^{\prime cc}\rightarrow Q^{\prime}and theorem 2 is proven.

It remains to demonstrate relations (5). These relations result from
I1I_{1}em me 1. The roots of the derivative of a polynomial whose roots are all real and simple, are functions that are increasing with respect to each of the roots of the polynomial.

Eithergga polynomial of degreen(1)n(\geqq 1)having all its rootsα1<α2<<αn\alpha_{1}<\alpha_{2}<<\ldots<\alpha_{n}real and simple. The property of Lemma 1 amounts to the fact that each of the roots of the derivative of the polynomialfhas=(xα)gf_{a}=(x-\alpha)gis an increasing function ofα\alphaThese rootsβ1<β2<<βn\beta_{1}<\beta_{2}<\ldots<\beta_{n}are continuous functions ofα\alphaand remain distinct. We can first demonstrate that they are strictly monotonic functions ofα\alphaIn effect, if, for example,βk\beta_{k}would not be a strictly monotonic function ofα\alphawe could find two different valuesα,α"\alpha^{\prime},\alpha^{\prime\prime}ofα\alphafor which the polynomials

fhas=(xα)g+g,fhas"=(xα")g+gf_{a^{\prime}}^{\prime}=\left(x-\alpha^{\prime}\right)g^{\prime}+g,\quad f_{a^{\prime\prime}}^{\prime}=\left(x-\alpha^{\prime\prime}\right)g^{\prime}+g (6)

have a common root equal toβk\beta_{k}This is impossible because any common root of the polynomials (6) would have to be a common root of the polynomialsg,gg,g^{\prime}which contradicts the hypothesis thatgghas only simple roots. The strict monotony of the rootsβi\beta_{i}functions ofα\alpha, is thus demonstrated. It only remains to specify the meaning of this monotonicity. If we note that the roots of the derivative are separated by those of the polynomial and if we take into account

limithashastβi=αi,i=1,2,,n\lim_{a\rightarrow a_{t}}\beta_{i}=\alpha_{i},\quad i=1,2,\ldots,n

we find it good that theβi\beta_{i}are increasing functions ofα\alpha
Lemma 1 is therefore proven. Other
proofs of Lemma 1 can be given. In particular, proofs based on considerations similar to those used in the second and third parts of this work can be given. We will not dwell on these proofs here.

Note. If

α1<α2<<αn1\alpha_{1}^{\prime}<\alpha_{2}^{\prime}<\ldots<\alpha_{n-1}^{\prime}

are the roots of the polynomialgg^{\prime}the rootsβi,i=1,2,,n\beta_{i},i=1,2,\ldots,nvary respectively within the intervals(,α1],[αi1,αi],i=2,3,,n1,[αn1,)\left(-\infty,\alpha_{1}^{\prime}\right],\left[\alpha_{i-1}^{\prime},\alpha_{i}^{\prime}\right],i=2,3,\ldots,n-1,\left[\alpha_{n-1}^{\prime},\infty\right), ifn>2n>2. Ifn=1n=1the rootβ1\beta_{1}varies from-\inftyhas\inftyand ifn=2n=2the rootsβ1\beta_{1},β2\beta_{2}vary respectively in the intervals(,α1+α22],[α1+α22,)\left(-\infty,\frac{\alpha_{1}+\alpha_{2}}{2}\right],\left[\frac{\alpha_{1}+\alpha_{2}}{2},\infty\right).

Go

  1. 5.

    We will now deal with the aforementioned proof of WA Markov's theorem.

We introduce the relationshipPsQP\xrightarrow{s}Qbetween two polynomials, which occurs if and only if:
11^{\circ}PolynomialsP,QP,Qhave the same degreen1n\geqq 1.
22^{\circ}The respective roots (1) of these polynomials satisfy the inequalities

x1y1x2y2xnynx_{1}\leqq y_{1}\leqq x_{2}\leqq y_{2}\leqq\cdots\leqq x_{n}\leqq y_{n} (7)

IfPsQP\xrightarrow{s}QOrQsPQ\xrightarrow{s}Pwe can say that the roots of the polynomialsP,QP,Qseparate. Generally, fromPsQP\xrightarrow{s}QIt follows thatPcQP\xrightarrow{c}Qand, forn=1n=1relationshipsPsQ,PcQP\xrightarrow{s}Q,P\xrightarrow{\mathrm{c}}Qare equivalent. Taking into account a previous remark, we can find, for alln>1n>1, two polynomialsP,QP,Qdegreennsuch that none of the relationshipsPSQP\xrightarrow{S}Q,QsPQ\xrightarrow{s}Pnot to be verified.

10. W. A. ​​Markov's theorem can be stated as follows:
THEOREM 3. IfP,QP,Qare two polynomials of degreen>1n>1, ofPsQP\xrightarrow{\mathrm{s}}Qit follows thatPsQP^{\prime}\xrightarrow{\mathrm{s}}Q^{\prime}.

Ifn=2n=2Theorem 3 follows from Theorem 1. Indeed, in this case, ofPsQP\xrightarrow{\mathrm{s}}Qit follows thatPcQP\xrightarrow{\mathrm{c}}QTherefore, by virtue of Theorem 1, it follows thatPcQP^{\prime}\xrightarrow{\mathrm{c}}Q^{\prime}But this relationship is equivalent toPsQP^{\prime}\xrightarrow{\mathrm{s}}Q^{\prime}and ownership is demonstrated.

We also introduce the relationshipPssQP\xrightarrow{ss}Qbetween two polynomials, which occurs if and only if:
11^{\circ}PolynomialsP,QP,Qhave the same degreen1n\geq 1and both have their very simple roots.

22^{\circ}The respective roots(1)\left(1^{\prime}\right)of these polynomials satisfy the inequalities

x1<y1<x2<y2<<xn<ynx_{1}<y_{1}<x_{2}<y_{2}<\ldots<x_{n}<y_{n} (\prime)

OfPssQP\xrightarrow{\mathrm{ss}}Qit follows thatPccQP\xrightarrow{\mathrm{cc}}Qand foru=1u=1These relationships are equivalent.

We have the following special case of W. A. ​​Markov's theorem:
THEOREM 4. IfP,QP,Qare two polynomials of degreen>1n>1, ofPssQP\xrightarrow{\mathrm{ss}}Qit follows thatPssQP^{\prime}\xrightarrow{\mathrm{ss}}Q^{\prime}.

It is demonstrated, as above, that forn=2n=2Theorem 4 follows from Theorem 2.

6. It suffices to prove Theorem 4, since Theorem 3 then follows from it. To see this, we proceed as in No. 3, where we showed that Theorem 1 follows from Theorem 2.
If we havePsQP\xrightarrow{\mathrm{s}}Qand if we now consider polynomialsPε,QεP_{\varepsilon},Q_{\varepsilon}having respectively as rootsxi+(2i1)ε,i=1,2,,nx_{i}+(2i-1)\varepsilon,i=1,2,\ldots,n,yi+2iε,i=1,2,,ny_{i}+2i\varepsilon,i=1,2,\ldots,n, Orε\varepsilonis a positive number, we havePεssQεP_{\varepsilon}\xrightarrow{\mathrm{ss}}Q_{\varepsilon}If we assume that Theorem 4 is true, it follows thatPεssQεP_{\varepsilon}^{\prime}\xrightarrow{\mathrm{ss}}Q_{\varepsilon}^{\prime}If we doε0\varepsilon\rightarrow 0, the roots ofPε,QεP_{\varepsilon}^{\prime},Q_{\varepsilon}^{\prime}tend towards the respective roots ofP,QP^{\prime},Q^{\prime}and we deduce thatPsQP^{\prime}\xrightarrow{\mathrm{s}}Q^{\prime}Theorem 3 is therefore proven.

7. We can prove Theorem 4, based on Theorem 2, Lemma 1 and the continuity of the roots of the derivative.

IfPssQP\xrightarrow{\mathrm{ss}}Qit follows thatPccQP\xrightarrow{\mathrm{cc}}Qso alsoPccQP^{\prime}\xrightarrow{\mathrm{cc}}Q^{\prime}To demonstrate that we have morePssQP^{\prime}\xrightarrow{\mathrm{ss}}Q^{\prime}It suffices to show that the derivatives of the polynomialsP,QP,Q(which all have simple roots) cannot have common roots. Indeed, it is easy to see that then the relationshipPssQP^{\prime}\xrightarrow{\mathrm{ss}}Q^{\prime}maintains itself while the roots ofPPgrow towards the respective roots ofQQ.

But, if(xi)(x_{i}^{\prime})are the roots of the polynomialsP,QP,Qdegreennthe relationshipPssQP\xrightarrow{\mathrm{ss}}Qis equivalent to equality

Q=P(has+i=1nhasixxi)Q=P\left(a+\sum_{i=1}^{n}\frac{a_{i}}{x-x_{i}}\right) (1)

Orhasais a non-zero constant andhas1,has2,,hasna_{1},a_{2},\ldots,a_{n}constants other than zero and of the same sign. Moreover, the producthashasiaa_{i}is of opposite sign with the highest coefficient ofPP, therefore withPPForxxvery large.

By derivation of (8) it follows that
(9)

Q=P(has+i=1nhasixxi)Pi=1nhasi(xxi)2Q^{\prime}=P^{\prime}\left(a+\sum_{i=1}^{n}\frac{a_{i}}{x-x_{i}}\right)-P\sum_{i=1}^{n}\frac{a_{i}}{(x-x_{i})^{2}}

We can see that ifP,QP^{\prime},Q^{\prime}had a common root, this root must also make the polynomial zeroPPwhich is impossible since, by hypothesis,PPhas only simple roots.

8. RelationshipsPsQP\xrightarrow{\mathrm{s}}QAndPssQP\xrightarrow{\mathrm{ss}}Qcan be extended to the case where the polynomialPPis of degreennand the polynomialQQdegreen1n-1. Ifn>1n>1and if

x1x2xn,y1y2yn1x_{1}\leq x_{2}\leq\ldots\leq x_{n},\quad y_{1}\leq y_{2}\leq\ldots\leq y_{n-1} (10)

are respectively the roots ofPPandQQthe relationshipPsQP\xrightarrow{\mathrm{s}}Qtakes place if and only if

xiyixi+1,i=1,2,,n1x_{i}\leq y_{i}\leq x_{i+1},\quad i=1,2,\ldots,n-1 (11)

We can also say that the roots of polynomialsP,QP,QThey separate.
The relationshipPssQP\xrightarrow{\mathrm{ss}}Qtakes place if and only if, in addition, the roots of the polynomialsP,QP,Qare all simple and if instead of (11) we have the inequalities

xi<yi<xi+1,i=1,2,,n1.x_{i}<y_{i}<x_{i+1},i=1,2,\ldots,n-1. (\prime)

We then have
Consequence 1. IfPPis a polynomial of degreennAndQQa polynomial of degreen1,n>1n-1,n>1, ofPsQP\xrightarrow{s}Qit follows thatPsQP^{\prime}\xrightarrow{s}Q^{\prime}.

The property follows from Theorem 3 by taking the limit. To see this, let (10) be the roots ofPPAndQQwhich verify the relationshipPsQP\xrightarrow{s}QConsider the polynomialR=(xyn)QR=\left(x-y_{n}\right)Qdegreenn. Ifxnynx_{n}\leqq y_{n}no we havePsRP\xrightarrow{\mathrm{~s}}R, hence, taking into account the theorem3,PsR3,P^{\prime}\xrightarrow{\mathrm{s}}R^{\prime}If we doy11y_{11}\rightarrow\infty, one of the roots ofRR^{\prime}(the largest) tends towards\inftyand the others towards the respective roots ofQQ^{\prime}Considering the continuity of the roots of the derivative with respect to the roots of the polynomial, we see that if we doyny_{n}\rightarrow\infty, ofPRP^{\prime}\rightarrow R^{\prime}it follows thatPQP^{\prime}\rightarrow Q^{\prime}.

We also have
Consequence 2. IfPPis a polynomial of degreennAndQQa polynomial of degreen1,n>1n-1,n>1, ofPssQP\xrightarrow{\mathrm{ss}}Qit follows thatPssQP^{\prime}\xrightarrow{\mathrm{ss}}Q^{\prime}.

This property follows from Theorem 4 in the same way as Consequence 1 of Theorem 3. We construct, as above, the polynomialRR. IfPssQP\xrightarrow{ss}Qand ifxn<ynx_{n}<y_{n}, We havePssRP\xrightarrow{ss}R, therefore, as a consequence of the theorem4,PsR4,P^{\prime}\xrightarrow{s}R^{\prime}From there, if we dojnj^{\prime}n\rightarrow\inftyit follows thatPQP^{\prime}\rightarrow Q^{\prime}To demonstrate that we evenPQP^{\prime}\rightarrow Q^{\prime}it suffices to demonstrate that ifPssQP\xrightarrow{ss}Q, polynomialsP,QP^{\prime},Q^{\prime}cannot have common roots.

IfPssQP\xrightarrow{\mathrm{ss}}Qwe have formula (8), wherehas=0a=0Andhasi,i=1,2,,na_{i},i=1,2,\ldots,nareuuconstants other than zero and of the same sign. Formula (9) shows us thatP,QP^{\prime},Q^{\prime}cannot have common roots.
La consequence 2L_{\text{a conséquence }}2is demonstrated.
9. As a function, consider a sequence of orthogonal polynomials

Pi0,Pi1,,Pin1,Pin,\Pi_{0},\Pi_{1},\ldots,\Pi_{n-1},\Pi_{n},\ldots

We know that the roots ofPin\Pi_{n}are all real, simple, and the roots ofPin1\Pi_{n-1}are separated in the strict sense by the roots ofPin\Pi_{n}So we havePinssPin1\Pi_{n}\xrightarrow{\mathrm{ss}}\Pi_{n-1}Forn>1n>1From consequence 2, it therefore follows that

Consequence 3. IfPin1,Pin\Pi_{n-1},\Pi_{n}are two consecutive terms of a sequence of orthogonal polynomials (n>1n>1), We havePinssPin1\Pi_{n}^{\prime}\xrightarrow{ss}\Pi_{n-1}^{\prime}.

III

  1. 10.

    We will deal with a theorem analogous to WA Markov's theorem.

We introduce the relationshipPmQP\xrightarrow{m}Qbetween two polynomials, which occurs if and only if:
1010^{\circ}PolynomialsP,QP,Qare of the same degreen1n\geqq 1.
22^{\circ}The roots (1) of these polynomials satisfy the inequalities
(12)x1+x2++xiy1+y2++yi,i=1,2,,n1\quad x_{1}+x_{2}+\ldots+x_{i}\leqq y_{1}+y_{2}+\ldots+y_{i},\quad i=1,2,\ldots,n-1
and equality

x1+x2++xn=y1+y2++ynx_{1}+x_{2}+\cdots+x_{n}=y_{1}+y_{2}+\cdots+y_{n} (13)

Ifn=1n=1, we only have equality (13). Ifn=1n=1the relationshipPmQP\xrightarrow{m}Qmeans thatP,QP,Qhave the same root, therefore, according to the meaning adopted at the beginning of this work, they are equal. It is easy to see that for alln>1n>1we can find two polynomialsP,QP,Qdegreennsuch that none of the relationshipsPmQ,QmPP\xrightarrow{\mathrm{~m}}Q,Q\xrightarrow{\mathrm{~m}}Pnot to be verified.

According to G.H. Hardy, J.E. Littlewood and G.P. Polya [1], the relationshipP ml QP\xrightarrow{\text{ ml }}Qis equivalent to the fact that the roots ofQQare deduced from those ofPPthrough a kind of "mediation" process. This means that there is a matrix (hasL,ja_{l,j}) hasnnlines andnncolumns, with non-negative elements, the sum of the elements in each row and each column being equal to 1,

v=1nhasi,v=v=1nhasv,j=1,i,j=1,2,,n\sum_{v=1}^{n}a_{i,v}=\sum_{v=1}^{n}a_{v,j}=1,\quad i,j=1,2,\ldots,n

and such that

yi=j=1nhasi,jxj,i=1,2,,ny_{i}=\sum_{j=1}^{n}a_{i,j}x_{j},i=1,2,\ldots,n

In what follows, we will not directly use this property.
The relation under consideration is (reflexive and) transitive, and we have the following theorem, analogous to W.A. Markov's theorem:

THEOREM 5. IfP,QP,Qare two polynomials of degreen>1n>1, ofPmQP\xrightarrow{\mathrm{m}}Qit follows thatPmQP^{\prime}\xrightarrow{\mathrm{m}}Q^{\prime}.

We also introduce the relationshipPmmQP\xrightarrow{\mathrm{mm}}Qbetween two polynomials, which occurs if and only if:
11^{\circ}PolynomialsP,QP,Qare of the same degreen1n\geq 1and all have simple roots.
22^{\circ}The respective roots of these polynomials satisfy the inequalities

x1+x2++xi<y1+y2++yi,i=1,2,,n1x_{1}+x_{2}+\ldots+x_{i}<y_{1}+y_{2}+\ldots+y_{i},\quad i=1,2,\ldots,n-1 (12')

and equality (13).
Forn=1n=1, we keep for the definition only equality (13) and then the relationPmmQP\xrightarrow{\mathrm{mm}}Qis equivalent toP111QP\xrightarrow{111}Q.

The relationshipmm\xrightarrow{\mathrm{mm}}is transitive. We also have mixed transitivity properties between the two relations considered. IfPmmRP\xrightarrow{\mathrm{mm}}R,RmQR\xrightarrow{\mathrm{m}}Qand ifQQto all its simple roots, we haveP1mnQP\xrightarrow{1mn}Q. OfP1mnRP\xrightarrow{1mn}R,RmSR\xrightarrow{\mathrm{m}}S,SmmQS\xrightarrow{\mathrm{mm}}Qit follows thatPmmQP\xrightarrow{\mathrm{mm}}Q.

Finally, we have:
THEOREM 6. IfP,QP,Qare two polynomials of degreen>1n>1, ofPm111QP\xrightarrow{\mathrm{m}111}Qit follows thatPm111QP^{\prime}\xrightarrow{\mathrm{m}111}Q^{\prime}.

The proof of theorems 5 and 6 is immediate forn=2n=2, since the root of the derivative of a polynomial of degree 2 is equal to half the sum of the roots of the polynomial. In this case, theorems 5 and 6 follow from equality (13).

There are two cases where the proof of Theorem 5 presents no difficulties. These cases occur if one of the polynomialsP,QP,Qto all its roots combined.

Let's start by making a few remarks. If

x1x2xnx_{1}\leq x_{2}\leq\ldots\leq x_{n} (14)

Ifx1,x2,,xnx_{1},x_{2},\ldots,x_{n}are the roots of the polynomialPP, We have

x1x1+x22x1+x2+x33x1+x2++xnn.x_{1}\leq\frac{x_{1}+x_{2}}{2}\leq\frac{x_{1}+x_{2}+x_{3}}{3}\leq\ldots\leq\frac{x_{1}+x_{2}+\ldots+x_{n}}{n}. (15)

Ifxix_{i}Andyiy_{i}are the roots of the polynomialsP,QP,Qand ifPmQP\xrightarrow{\mathrm{m}}Q, We havex1y1x_{1}\leq y_{1},ynxny_{n}\leq x_{n}, therefore also

xnx1yny10.x_{n}-x_{1}\geq y_{n}-y_{1}\geq 0. (16)

Ifn>1n>1AndP11111QP\xrightarrow{11111}QWe have more precise inequalities

xnx1>yny1>0.x_{n}-x_{1}>y_{n}-y_{1}>0.

Let us now demonstrate Theorem 5 in the two specific cases mentioned:

Case 1. The polynomialPPto all its roots combined. OfP11QP\xrightarrow{11}Qand (16), it follows thatQQalso includes all its roots together, and in particular the single distinct root ofPP. In this casePP^{\prime}AndQQ^{\prime}also have all their roots conflated with the single distinct root ofPP, and theorem 5 follows from this.

Case 2. The polynomialQQto all its roots combined. Beξi,ηi\xi_{i},\eta_{i}the roots of polynomialsP,QP^{\prime},Q^{\prime}and let us take into account the inequalities (15) corresponding to these roots. Then ifP111QP\xrightarrow{111}Q, We have

ξ1ξ1+ξ22ξ1+ξ2++ξn1n1=η1=η2==ηn1,\xi_{1}\leq\frac{\xi_{1}+\xi_{2}}{2}\leq\ldots\leq\frac{\xi_{1}+\xi_{2}+\ldots+\xi_{n-1}}{n-1}=\eta_{1}=\eta_{2}=\ldots=\eta_{n-1},

from which it follows thatPQP^{\prime}\xrightarrow{}Q^{\prime}and Theorem 5 is proven.

To go further, we introduce two operations on the roots of a polynomial. We will call these operations: dilation and contraction of two roots. These operations have already been used by G.H. Hardy, J.F. Littlewood, and G. Pólya.

A dilation of two of the rootsxx"x^{\prime}\leq x^{\prime\prime}of a polynomial amounts to substituting these roots byxρx^{\prime}-\rho,x"+ρx^{\prime\prime}+\rhorespectively, whereρ>0\rho>0and the other roots of the polynomial remain unchanged.

A contraction of two of the rootsx<x"x^{\prime}<x^{\prime\prime}of a polynomial amounts to substituting these roots byx+ρx^{\prime}+\rho,x"ρx^{\prime\prime}-\rhorespectively, whereρ>0\rho>0and the other roots of the polynomial remain unchanged.

The numberρ\rhocan be called the coefficient of dilation or contraction corresponding to the pair of roots considered.

In what follows, unless expressly stated otherwise, we only consider dilations and contractions that do not disturb the order of the polynomial's roots. This means that the coefficient is subject to the restriction that, in the case of dilation, the intervals[xρ,x)[x^{\prime}-\rho,x^{\prime}),(x",x"+ρ](x^{\prime\prime},x^{\prime\prime}+\rho]and in the case of contraction, the intervals[x,x+ρ][x^{\prime},x^{\prime}+\rho],[x"ρ,x"][x^{\prime\prime}-\rho,x^{\prime\prime}]contain no roots of the original polynomial or the transformed polynomial. Ifxix_{i}are the roots of the initial polynomial and ifn>1n>1With the previous restriction, the dilation operation is applicable to the rootsxr,xsx_{r},x_{s},r<sr<sonly in the following cases:

r=\displaystyle r= 1,s=n, For 0<ρ any ,\displaystyle 1,s=n,\text{ pour }0<\rho\text{ quelconque },
r=\displaystyle r= 1,s<n, if x1xs<xs+1, For 0<ρ<xs+1xs,\displaystyle 1,s<n,\text{ si }x_{1}\leqq x_{s}<x_{s+1},\text{ pour }0<\rho<x_{s+1}-x_{s},
r>\displaystyle r> 1,s=n, if xr1<xrxn, For 0<ρ<xrxr1,\displaystyle 1,s=n,\text{ si }x_{r-1}<x_{r}\leqq x_{n},\text{ pour }0<\rho<x_{r}-x_{r-1},
r>\displaystyle r> 1,s<n, if xr1<xrxs<xs+1, For\displaystyle 1,s<n,\text{ si }x_{r-1}<x_{r}\leqq x_{s}<x_{s+1},\text{ pour }
0<ρ<min(xrxr1,xs+1xs).\displaystyle 0<\rho<\min\left(x_{r}-x_{r-1},x_{s+1}-x_{s}\right).

Similarly, the contraction operation is applicable only in the following cases:

sγ=1, if xr<xr+1, For 0<ρ<xr+1xr2,s-\gamma=1,\text{ si }x_{r}<x_{r+1},\text{ pour }0<\rho<\frac{x_{r+1}-x_{r}}{2},

sr>1s-r>1, ifxr<xr+1xs1<xsx_{r}<x_{r+1}\leqq x_{s-1}<x_{s}, For0<ρ<min(xr+1xr,xsxs1)0<\rho<\min\left(x_{r+1}-x_{r},x_{s}-x_{s-1}\right)
With the dilation and contraction operations thus specified, we see that such an operation is perfectly characterized by the pair of roots considered and the coefficient ρ\rhocorresponding. In particular, if we can apply a dilation or contraction operation of the coefficientρ\rhoWe can also apply to the same roots any dilation or contraction of coefficient<ρ<\rho.

We deduce that if we apply to the rootsxr,xs,r<sx_{r},x_{s},r<sIn the case of a polynomial, whether it's a dilation or a contraction, the roots of the polynomial become continuous functions of the coefficientρ\rhoThe same applies to sumsx1+x2++xi,i=1,2,,nx_{1}+x_{2}+\ldots+x_{i},i=1,2,\ldots,nThese sums are transformed intox1+x2++xiρx_{1}+x_{2}+\ldots+x_{i}-\rhorespectively inx1+x2++xi+ρx_{1}+x_{2}+\ldots+x_{i}+\rhoForı=r,r+1,,s1\imath=r,r+1,\ldots,s-1depending on whether it is a dilation or a contraction of coefficientρ\rhorootsxr,xs,r<sx_{r},x_{s},r<sThe sumsx1+x2++xLx_{1}+x_{2}+\ldots+x_{l}, for the other values ​​ofiiremain invariable. It is important to remember, in particular, that the sum of all the roots remains invariable under a dilation or contraction of two roots.

IfP*P^{*}is a polynomial that is derived from the polynomialPPby applying a dilation or contraction coefficientρ\rho, the roots ofP*P^{*}tend, forρ0\rho\rightarrow 0, towards the corresponding roots ofPPAt the same time, the roots ofP*P^{*\prime}, which are also continuous functions ofPP, tend towards the corresponding roots ofPP^{\prime}.

It is important to extend these properties to the limit, in caseP*P^{*}can be deduced fromPPby successively applying a finite number of expansions or contractions relative to various complexes of roots of the polynomial. This extension must be done with certain precautions because the successive application of several operations depends on their order. The
expansion and contraction operations are therefore not commutative when applied to pairs of different roots.

Example. Letn=3n=3Andx1=0,x2=x3=2x_{1}=0,x_{2}=x_{3}=2The first root is therefore equal to 0, the second and third to 2. If we first apply to the rootsx1,x3x_{1},x_{3}(at the first and third) a dilation with a coefficient of 3, the roots become3,2,5-3,2,5Then, by applying a contraction with a coefficient of 1 to the rootsx2,x3x_{2},x_{3}(in the second and third steps) we obtain the roots3,3,4-3,3,4The order of operations cannot be reversed because the contraction operation cannot be applied to roots.x2=2,x3=2x_{2}=2,x_{3}=2, if we take into account the imposed restriction of not disturbing the order of the roots.

In this case, let's assume we first apply it to the rootsx2,x3x_{2},x_{3}(in the second and third) a contraction with a coefficient of 1. The roots become0,1,30,1,3We then apply a dilation with a coefficient of 3 to the roots 0 and 1 (the first and second). We recover the roots3,3,4-3,3,4It is worth noting that each time we have disrupted the order of the roots.

This example illustrates, firstly, how the restriction of not disturbing the order of the roots clarifies the operations of expansion and contraction. Secondly, the same example shows us how to follow the roots of the polynomial when successively applying several expansions or contractions of two roots.

We will not dwell on this permutability problem, as the boundary properties from above will only be applied in the following cases, which will be specified when they actually arise.
13. We will now demonstrate that Theorem 5 follows from Theorem 6.

If the polynomialP*P^{*}is obtained from the polynomialPPBy applying a dilation to two roots, we haveP*mPP^{*}\xrightarrow{m}PThis relationship is also true when the polynomialP*P^{*}is obtained fromPPby the successive application of any number of expansions.

Let (14) be the roots of the polynomialPPand eithern>1n>1Let us designate byPeP_{e}a polynomial whose roots are the numbers

xi=xi(ni)p,i=1,2,,n1,xn=xn+n(n1)2ρ,x_{i}^{\prime}=x_{i}-(n-i)p,i=1,2,\ldots,n-1,x_{n}^{\prime}=x_{n}+\frac{n(n-1)}{2}\rho,

Orρ\rhois a positive number. The polynomialPρP_{\rho}is obtained fromPPby successively applying the operations of dilation, of coefficient(ni)ρ(n-i)\rhoto the rootsxi,xnx_{i},x_{n}, Fori=1,2,,n1i=1,2,\ldots,n-1(in that order). We have Pe11PP_{\mathrm{e}}\xrightarrow{11}PNote thatPeP_{\mathrm{e}}has all its simple roots which, forρ0\rho\rightarrow 0, tend towards the corresponding roots ofPPAt the same time, the roots ofPeP_{e}^{\prime}tend towards the corresponding roots ofPP^{\prime}.

LetP,QP,Qtwo polynomials of degreen>1n>1, (1) the roots of these polynomials and suppose thatPmQP\xrightarrow{m}QLet

x1<x2<<xn,y1<y2<<ynx_{1}^{\prime}<x_{2}^{\prime}<\ldots<x_{n}^{\prime},\quad y_{1}^{\prime}<y_{2}^{\prime}<\ldots<y_{n}^{\prime}

the roots of polynomialsP2ϱ,QeP_{2\varrho},Q_{e}, Orρ\rhois a positive number and which are obtained fromP,QP,Qin the same way as above the polynomialPQP_{Q}ofPPWe then have

y1+y2++yi(x1+x2++xi)==y1+y2++yi(x1+x2++xi)+i(2ni1)2ρ>0,i=1,2,,n1,x1+x2++xn=y1+y2++yn.\begin{gathered}y_{1}^{\prime}+y_{2}^{\prime}+\ldots+y_{i}^{\prime}-\left(x_{1}^{\prime}+x_{2}^{\prime}+\ldots+x_{i}^{\prime}\right)=\\ =y_{1}+y_{2}+\ldots+y_{i}-\left(x_{1}+x_{2}+\ldots+x_{i}\right)+\frac{i(2n-i-1)}{2}\rho>0,\\ i=1,2,\ldots,n-1,\\ x_{1}^{\prime}+x_{2}^{\prime}+\ldots+x_{n}^{\prime}=y_{1}^{\prime}+y_{2}^{\prime}+\ldots+y_{n}^{\prime}.\end{gathered}

So we haveP2ϱmmQeP_{2\varrho}\xrightarrow{mm}Q_{e}Assuming, therefore, that Theorem 6 is true, it follows thatP2ρmmQϱP_{2_{\rho}}^{\prime}\xrightarrow{mm}Q_{\varrho}^{\prime}But, ifρ0\rho\rightarrow 0, the roots ofP2ρ,QϱP_{2_{\rho}}^{\prime},Q_{\varrho}^{\prime}tend towards the roots ofP,QP^{\prime},Q^{\prime}respectively. So, the end is thusp0p\rightarrow 0, we deduce thatP1mQP^{\prime}\xrightarrow{1m}Q^{\prime}.

We have thus demonstrated that Theorem 5 follows from Theorem 6.
Note: The relationP*111PP^{*}\xrightarrow{111}Pis true ifP*P^{*}is obtained fromPPby dilating two roots, without the restriction of preserving the root order. Indeed, this is easily seen if we apply a dilation to the rootsxx"x^{\prime}\leq x^{\prime\prime}and if we assume that the coefficientρ\rhobelieves, we can replacexρx^{\prime}-\rhooutx"ρx^{\prime\prime}-\rhothrough a root that it passes through. Let's agree that a dilation of the rootsxx"x^{\prime}\leq x^{\prime\prime}does not broadly affect the order of the roots when the intervals (xρ,xx^{\prime}-\rho,x^{\prime}), (x",x"+ρx^{\prime\prime},x^{\prime\prime}+\rho) do not contain roots of the polynomial. Then the previous property follows from the fact that any dilation, without the restriction of the order of the roots, is obtained by the successive application of a finite number of dilations which does not disturb the order of the roots in a broad sense.

We also see that the relationshipP*11PP^{*}\xrightarrow{11}Pis true when we obtainP*P^{*}ofPPby the successive application of any number (finite or not) of dilations, with or without the restriction of preserving the order of the roots.
14. We now propose to prove Theorem 6. We will deduce it from a series of preliminary lemmas.

When the polynomialP*P^{*}is obtained fromPPby a contraction of two roots, we havePP*P\rightarrow P^{*}This relationship remains true ifP*P^{*}is obtained from
PPby a succession of a finite number of contractions. If the polynomialPPThe same is true for the polynomial.P*P^{*}.

Lemma 2. Given a polynomialPPdegreen>1n>1and for any positive number a, we can find a polynomial of degreennsuch as:11^{\circ}This polynomial can be deduced fromPPby the successive application of a finite number of contractions of two consecutive roots,22^{\circ}The roots of this polynomial must all be contained within an interval of length<ε<\varepsilon.

It is clear that ifPPConsidering all its roots, we have nothing to prove. Here we only consider the case wherePPhas all its simple roots, which we will assume later. It is also clear that the lemma remains true even without this restriction.

The statement specifies that it only deals with contractions applied to consecutive roots; therefore, if (14) are the roots of the polynomial, then only to pairs of roots of the formxi,xi+1x_{i},x_{i+1}.

We will prove the lemma by complete induction.
Forn=2n=2the property is true because ifx1<x2x_{1}<x_{2}are the roots ofPP, it suffices to apply a coefficient contraction to themρ\rhowho verifies inequalitiesmax(0,x2x1ε2)<ρ<x2x12\max\left(0,\frac{x_{2}-x_{1}-\varepsilon}{2}\right)<\rho<\frac{x_{2}-x_{1}}{2}

Let us now suppose thatn>2n>2and that the property is true for polynomials of degreen1n-1Let us demonstrate that the property will also hold true for polynomials of degreenn.

We will first demonstrate that ifPPis a polynomial of degreennhaving the roots (14)(x1<xn)\left(x_{1}<x_{n}\right)By applying a finite number of consecutive root contractions, we can deduce a polynode whose roots lie within an interval of length<xnx12+ε4<\frac{x_{n}-x_{1}}{2}+\frac{\varepsilon}{4}Indeed
, by hypothesis, by applying a finite number of consecutive root contractions, we can deduce fromPPa polynomialP1P_{1}including: the rootsx1<x2<<xnx_{1}^{\prime}<x_{2}^{\prime}<\ldots<x_{n}^{\prime}verify the relationshipsx1=x1,xnx2<84x_{1}^{\prime}=x_{1},x_{n}^{\prime}-x_{2}^{\prime}<\frac{8}{4}Contractions are applied only to couples of the formxi,xi+1x_{i},x_{i+1}, Ori>1i>1Next, we apply it to the polynomial.P1P_{1}a contraction of the rootsx1,x2x_{1}^{\prime},x_{2}^{\prime}coefficientρ\rhowhere max(0,x2x12ε8)<ρ<x2x12\left(0,\frac{x_{2}^{\prime}-x_{1}^{\prime}}{2}-\frac{\varepsilon}{8}\right)<\rho<\frac{x_{2}^{\prime}-x_{1}^{\prime}}{2}The roots of the polynomial thus obtained are then contained within an interval of length<xn(x1+ρ)<xnx1x2x12+ε3=xnx12++x2x12+ε8<xnx12+ε4<x_{n}^{\prime}-\left(x_{1}+\rho\right)<x_{n}^{\prime}-x_{1}^{\prime}-\frac{x_{2}^{\prime}-x_{1}^{\prime}}{2}+\frac{\varepsilon}{3}=\frac{x_{n}^{\prime}-x_{1}^{\prime}}{2}++\frac{x_{2}^{\prime}-x_{1}^{\prime}}{2}+\frac{\varepsilon}{8}<\frac{x_{n}-x_{1}}{2}+\frac{\varepsilon}{4}that's precisely what needed to be demonstrated.

It follows that if a polynomial of degreennhas all its roots contained within an interval of length<L<lBy applying a finite number
of consecutive root contractions, we can deduce a polymoma whose roots lie within an interval of length<L2+ε4<\frac{l}{2}+\frac{\varepsilon}{4}By repeating this process, we see that, for every natural numberkkWe can deduce, by the successive application of a finite number of contractions of two consecutive roots, a polynomial of degreennwhose roots are contained within an interval of length smaller than

L2k+ε(122+123++12k+1)<L2k+ε2\frac{l}{2^{k}}+\varepsilon\left(\frac{1}{2^{2}}+\frac{1}{2^{3}}+\ldots+\frac{1}{2^{k+1}}\right)<\frac{l}{2^{k}}+\frac{\varepsilon}{2}

Simply choose the numberkkso thatL2k<ε2\frac{l}{2^{k}}<\frac{\varepsilon}{2}and the lemma is, Clémontré.

Note. A similar observation can be made to the one made in point 13. The relationPmP*P\xrightarrow{m}P^{*}is also true whenP*P^{*}is obtained fromPPby the contraction of two tacinesx<x"x^{\prime}<x^{\prime\prime}, without the restriction of preserving the order of the roots, but with the condition that the coefficientρ\rhoeither<x"x<x^{\prime\prime}-x^{\prime}The demonstration is analogous, replacingx+ρx^{\prime}+\rhoOrx"ρx^{\prime\prime}-\rhowith every root it passes through, and in particular, by permuting these roots as they intersect, whileppbelieves. Here again we can agree to say that the rootsx<x"x^{\prime}<x^{\prime\prime}do not disturb, in a broad sense, the order of the roots when the intervals (x,x+ρx^{\prime},x^{\prime}+\rho), (x"ρ,x"x^{\prime\prime}-\rho,x^{\prime\prime}) do not contain any roots of the transformed polynomial and when0<ρ<x"x0<\rho<x^{\prime\prime}-x^{\prime}Then the previous property follows from the fact that any contraction of two roots, with the sole restriction that its coefficientρ\rhochecks inequalities0<ρ<x"x0<\rho<x^{\prime\prime}-x^{\prime}, can be obtained by a finite number of successive contractions which do not broadly disturb the order of the roots.

We also see that the relationshipPmP*P\xrightarrow{m}P^{*}is true whenP*P^{*}is obtained fromPPby the successive application of any number (finite or unlimited) of contractions with preservation of the order of the roots, or only with the restriction imposed above on the coefficients of the contractions.
15. From the preceding lemma it follows that
I1\mathrm{I}_{1}emme 3. IfP,QP,Qare two polynomials of degreen>2n>2and ifP1111QP\xrightarrow{1111}Qwe can find a polynomialRRof degree n, which is obtained fromPPby the successive application of a finite number of contractions of two consecutive roots, such that one hasRmQR\xrightarrow{\mathrm{~m}}Q, without the relationshipRmmQR\xrightarrow{\mathrm{~mm}}Qbe verified.

Let (11^{\prime}) the roots of polynomialsP,QP,QAndz1<z2<<znz_{1}<z_{2}<\ldots<z_{n}the roots ofRR. We haveP11RP\xrightarrow{11}Rand, by virtue of the assumptions verified byRR, 11we have inequalities
(17)z1+z2++ziy1+y2++yi,i=1,2,,n1z_{1}+z_{2}+\cdots+z_{i}\leqq y_{1}+y_{2}+\cdots+y_{i},\quad i=1,2,\ldots,n-1In at least one of them ,
the equality relation is true. Of course, we also havez1+z2++zn=y1+y2++ynz_{1}+z_{2}+\ldots+z_{n}=y_{1}+y_{2}+\ldots+y_{n}.

To demonstrate the lemma, let us take a positive number a such that

ε<yny1\varepsilon<y_{n}-y_{1} (18)

By Lemma 2, we can find a finite sequence of polynomials of degreenn,

P0,P1,,PkP_{0},P_{1},\ldots,P_{k}

such as:
11{}^{\circ}Each termPiP_{i}is obtained from the previousPi1P_{i-1}by a contraction of two consecutive roots.
22^{\circ}The first termP0P_{0}is equal toPPand the dermicrPkP_{k}has all its roots contained within an interval of length<ε<\varepsilon.

By hypothesisP0111mQP_{0}\xrightarrow{111m}QTherefore, there is a greater index.γ\gammasuch asPr mml QP_{r}\xrightarrow{\text{ mml }}QWe cannot haver=kr=k, because otherwise inequality (18) would contradict inequalities(16)\left(16^{\prime}\right)So we haveγ<k\gamma<k, therefore alsor+1kr+1\leqq kand the polynomialPr+1P_{r+1}does not verify the relationshipPr+1mmQP_{r+1}\xrightarrow{mm}Q. Eitherρ1\rho_{1}the contraction coefficient by whichPr+1P_{r+1}is obtained fromPrP_{r}. EitherP*P^{*}a polynomial that is obtained fromPPby applying to the same pair of (consecutive) roots a contraction of coefficientpp1p\leqq p_{1}. Whenρ0\rho\rightarrow 0, the roots ofP*P^{*}tend towards the corresponding roots ofPPr and whenρρ1\rho\rightarrow\rho_{1}they tend towards the corresponding roots of Perrr_{r}hompleρρ1\rho\rightarrow\rho_{1}hes desPr+1P_{r+1}. End by virtue of continuity in relation toρ\rhoof the roots, there are 111 positive numbersρρ1\rho\leqq\rho_{1}such as one hasP**111QP^{**}\xrightarrow{111}Q, without the relationshipP*mmQP^{*}\xrightarrow{mm}Qbe verified. Ein taking the polynomialRRequal to the polynomialP*P^{*}corresponding to thisρ\rho, the lemma is proven.
16. We also have the
I1\mathrm{I}_{1}em me 4 . SiP,QP,Qare two polynomials of degreen>1n>1and ifL11111Ql\xrightarrow{11111}Q, we can find a finite sequence of polynomials of degreenn,

P0,P1,,PkP_{0},P_{1},\ldots,P_{k} (19)

such as:
11^{\circ}Each termPiP_{i}is obtained from the preceding tearPi1P_{i-1}by a contraction of two consecutive roots.
22^{\circ}The first term is equal toPPct the last term is equal toQQWe
can denounce it by complete induction.
From contractions of consecutive roots, we can deduce a polynomial whose roots lie within an interval of length<L2+ε4<\frac{l}{2}+\frac{\varepsilon}{4}By repeating this process, we see that, for any natural numberkkWe can deduce, by the successive application of a finite number of contractions of two consecutive roots, a polynomial of degreennwhose roots are contained within an interval of length smaller than

L2k+ε(122+123++12k+1)<L2k+ε2\frac{l}{2^{k}}+\varepsilon\left(\frac{1}{2^{2}}+\frac{1}{2^{3}}+\cdots+\frac{1}{2^{k+1}}\right)<\frac{l}{2^{k}}+\frac{\varepsilon}{2}

Simply choose the numberkkso thatL2k<ε2\frac{l}{2^{k}}<\frac{\varepsilon}{2}and the lemma is proven.

Note. A similar observation can be made to that made in no. 13. The relationshipPmP*P\xrightarrow{\boldsymbol{m}}P^{*}is also true whenP*P^{*}is obtained fromPPby a contraction of two rootsx<x"x^{\prime}<x^{\prime\prime}, without the restriction of preserving the order of the roots, but with the condition that the coefficientρ\rhoeither<x"x<x^{\prime\prime}-x^{\prime}The demonstration is analogous, replacingx+ρx^{\prime}+\rhoOrx"ρx^{\prime\prime}-\rhowith every root it passes through, and in particular, by permuting these roots as they intersect, whileρ\rhogrows. Here again we can agree to say that the rootsx<x"x^{\prime}<x^{\prime\prime}do not broadly disrupt the order of the roots when the intervals(x,x+ρ),(x"ρ,x")\left(x^{\prime},x^{\prime}+\rho\right),\left(x^{\prime\prime}-\rho,x^{\prime\prime}\right)do not contain any roots of the plyome tar p, (ρ\rho. Then the preceding property results from the fact that any contraction of two roots, with the sole restriction that its coefficientρ\rhochecks inequalities0<ρ<x"x0<\rho<x^{\prime\prime}-x^{\prime}, can be obtained by a finite number of successive contractions which do not broadly disturb the order of the roots.

We also see that the relationshipPmP*kP^{m}P^{*k}is true whenP*P^{*}is obtained fromPPby the successive application of any number (finite or unlimited) of contractions with preservation of the order of the roots, or only with the restriction imposed above on the coefficients of the contractions.
15. From the preceding lemma it follows that
I1\mathrm{I}_{1}e mme 3 . SiP,QP,Qare two polynomials of degreen>2n>2and ifP mm QP\xrightarrow{\text{ mm }}Qwe can find a polynomialRRof degreen, which is obtained fromPPby the successive application of a finite number of contractions of two consecutive roots, such that one hasRmQR\xrightarrow{m}Q, without the relationshipRmmQR\xrightarrow{mm}Qbe verified.

Let (11^{\prime}) the roots of polynomialsP,QP,QAndz1<z2<<znz_{1}<z_{2}<\ldots<z_{n}the roots ofRR. We haveP1nRP\xrightarrow{1n}Rand, by virtue of the assumptions verified byRR, we have inequalities
(17)z1+z2++ziy1+y2++yi,i=1,2,,n1z_{1}+z_{2}+\ldots+z_{i}\leqq y_{1}+y_{2}+\ldots+y_{i},\quad i=1,2,\ldots,n-1In one case ,
the equality relation is true. Of course, we also havez1+z2++zn=y1+y2++ynz_{1}+z_{2}+\ldots+z_{n}=y_{1}+y_{2}+\ldots+y_{n}.

To demonstrate the lemma, let's take a positive numberε\varepsilonsuch as

ε<yny1.\varepsilon<y_{n}-y_{1}. (18)

By Lemma 2, we can find a finite sequence of polynomials of degreenn,

P0,P1,,PkP_{0},P_{1},\ldots,P_{k}

such as:
11^{\circ}Each termPiP_{i}is obtained from the previousPi1P_{i-1}by a contraction of two consecutive roots.
22^{\circ}The first termP0P_{0}is equal toPPand the last onePkP_{k}has all its roots contained within an interval of length<ε<\varepsilon.

By hypothesisP0QP_{0}\rightarrow QTherefore, there is a greater index.rrsuch asPr mmll QP_{r}\xrightarrow{\text{ mmll }}QWe cannot haver=kr=k, because otherwise inequality (18) would be in contradiction with inequalities (1616^{\prime}). So we haveγ<k\gamma<k, therefore alsor+1kr+1\leqq kand the polynomialPr+1P_{r+1}does not check the relationshipPr+1mm0P_{r+1}\xrightarrow{mm}0. Eitherρ1\rho_{1}the contraction coefficient by whichPr+1P_{r+1}is obtained fromPrP_{r}. EitherP*P^{*}a polynomial that is obtained fromPPby applying to the same pair of (consecutive) roots a contraction of coefficientρρ1\rho\leqq\rho_{1}. Whenp0p\rightarrow 0, the roots ofP*P^{*}tend towards the corresponding roots ofPrP_{r}and whenρρ1\rho\rightarrow\rho_{1}they tend towards the corresponding roots ofPr+1P_{r+1}By virtue of continuity with respect toρ\rhoof the roots, there are 111 positive numbersρρ1\rho\leqq\rho_{1}such asLl^{\prime}we haveP*mQP^{*}\xrightarrow{m}Q, without the relationshipP*mmQP^{*}\xrightarrow{mm}Qbe verified. By taking the polynomialRRequal to the polynomialP*P^{*}corresponding to thisρ\rho, the lemma is proven.
16. We also have the

I, em me 4. SiP,QP,Qare two polynomials of degreen>1n>1and ifP sill QP\xrightarrow{\text{ silll }}Q, we can find a finite sequence of polynomials of degreenn,

P0,P1,,PkP_{0},P_{1},\ldots,P_{k} (19)

such as:
11^{\circ}Each termPiP_{i}is obtained from the preceding termPi1P_{i-1}by a conclusion of two consecutive roots.
22^{\circ}The first term is equal toPPwhere the last term is equal toQQWe
can demonstrate this by full induction.

Forn=2n=2all you have to do is takek=1k=1, SOP0=P,P1=QP_{0}=P,P_{1}=Qand the first lemma is proven.

Let's taken>2n>2and suppose that the property is true for polynomials of degree2,3,,n12,3,\ldots,n-1Let us demonstrate that it will also be true for polynomials of degreenn.

Let us therefore consider two polynomialsP,QP,Qdegreennand suppose thatPmmQP\xrightarrow{mm}QBy virtue of Lemma 3 we can construct a finite sequence of polynomials of degreenn,

P0,P1,,Pr,RP_{0},P_{1},\ldots,P_{r},R (20)

OrP0P_{0}is equal toPPand whose terms satisfy condition 1 of Lemma 4. Furthermore, the last termRR, determined by lemma 3 , satisfies the relationRmQR\xrightarrow{\mathrm{~m}}Qbut no, not the relationshipRmn1QR\xrightarrow{\mathrm{~m}n1}QWe will continue to refer to byz1<z2<<znz_{1}<z_{2}<\ldots<z_{n}the roots ofRR.

IfRRis equal toQQ, sequence (20) satisfies all the conditions imposed on sequence (19) and lemma 4 is proven.

Otherwise, onlyjj, Or1j<n11\leqq j<n-1, relations (17) reduce to equalities. Leti1,i2,,iji_{1},i_{2},\ldots,i_{j}the values ​​ofiifor which we have equality in (17), for the other values ​​ofiiThe strict inequality (i.e., with < ) is valid. We can assume0=i0<i1<i2<<ij<ij+1=n0=i_{0}<i_{1}<i_{2}<\ldots<i_{j}<i_{j+1}=nLet us now consider pairs of consecutive indices.is,is+1i_{s},i_{s+1}These couples fall into two categories:
1Ifis+1is=11^{\circ}\mathrm{Si}i_{s+1}-i_{s}=1The couples are in the first category and we have 11zis+1=yis+1z_{i_{s}+1}=y_{i_{s}+1}.
2Ifis+1is>12^{\circ}\mathrm{Si}i_{s+1}-i_{s}>1The couples are in the second category, and in this case we have

zis+1+zis+2++zis+v<yis+1+yis+2++yis+vv=1,2,,is+1is1zis+1+zis+2++zis+1=yis+1+yis+2++yis+1\begin{gathered}z_{i_{s}+1}+z_{i_{s}+2}+\ldots+z_{i_{s}+v}<y_{i_{s}+1}+y_{i_{s}+2}+\ldots+y_{i_{s}+v}\\ v=1,2,\ldots,i_{s+1}-i_{s}-1\\ z_{i_{s}+1}+z_{i_{s}+2}+\therefore+z_{i_{s+1}}=y_{i_{s}+1}+y_{i_{s}+2}+\ldots+y_{i_{s+1}}\end{gathered}

But we haveis+1is<ni_{s+1}-i_{s}<nand, by virtue of the assumptions made, Lemma 4 is true for polynomials of degree<n<nIt follows that we can successively apply toRRa finite number of contractions of two consecutive rootszi,zi+1z_{i},z_{i+1}, Oris+1iis+11i_{s}+1\leqq i\leqq i_{s+1}-1so that the rootszis+1,zis+2,,zis1z_{i_{s}+1},z_{i_{s}+2},\ldots,z_{i_{s-1}}become respectively equal toyis+1,yis+2,,yis+1y_{i_{s}+1},y_{i_{s}+2},\ldots,y_{i_{s+1}}leaving the other roots unchanged. It follows, therefore, that we can extend the sequence (20) such that

P0,P1,,Pr,R,R1,,RrP_{0},P_{1},\ldots,P_{r},R,R_{1},\ldots,R_{r^{\prime}}

where the terms satisfy the same conditions as those of sequence (20), except that the last termRrR_{r^{\prime}}, has a number of one less unit of consecutive index pairs of the second category.

Since, obviously, only a finite number of pairs of consecutive indicesis,is+1i_{s},i_{s+1}of the second category exist, we see that, by possibly repeating a finite number of times the process above, we manage to construct a sequence (19), by suitably extending the sequence (20) and which satisfies all the conditions of lemma 4.

Lemma 4 is therefore proven.
17. Finally, we have the
I1\mathrm{I}_{1}emme 5. IfPPis a polynomial of degreen>1n>1, having all its simple roots and if the polynomialQQis obtained fromPPby a contraction of two consecutive roots, we havePmmQP^{\prime}\xrightarrow{mm}Q^{\prime}.

Before proving this lemma, we will show that Theorem 6 then follows from it. Indeed, letP,QP,Qtwo polynomials of degreen>1n>1and suppose thatPmmQP\xrightarrow{mm}QWe apply Lemma 4, constructing a sequence (19) which satisfies the properties1,21^{\circ},2^{\circ}of this lemma. By virtue of Lemma 5, we then havePi1111mPi,i=1,2,,kP_{i-1}^{\prime}\xrightarrow{111m}P_{i}^{\prime},i=1,2,\ldots,k, hence, taking into account the transitivity of the relationmm\xrightarrow{mm}, we deduceP0mmPkP_{0}^{\prime}\xrightarrow{mm}P_{k}^{\prime}, therefore alsoPmmQP^{\prime}\xrightarrow{mm}Q^{\prime}and theorem (i) is proven.
18. It remains to prove lemma 5.

From the above, it follows that it suffices to demonstrate thatn>2n>2It is then easy to see that lemma 5 is equivalent to

Lemma 6. Ifhas<b,0<ρ<bhas2a<b,0<\rho<\frac{b-a}{2}and iff=(xhas)(xb)h,g==(xhasρ)(xb+ρ)hf=(x-a)(x-b)h,g==(x-a-\rho)(x-b+\rho)h, Orhhis a polynomial of degreen>0n>0having all its roots real, simple, and located outside the closed interval[has,b][a,b], We havef"mmgf^{\prime\prime}\xrightarrow{mm}g^{\prime}.

The polynomialggresults from the polynomialffby applying a contraction to the rootshasaAndbb.

Let us designate by

x1<x2<<xn,y1<y2<<yn1x_{1}<x_{2}<\ldots<x_{n},\quad y_{1}<y_{2}<\ldots<y_{n-1}

the roots ofhhAndhh^{\prime}(ifn>1n>1) and let us designate by

z1<z2<<zn+1,z1<z2<<zn+1z_{1}<z_{2}<\ldots<z_{n+1},z_{1}^{\prime}<z_{2}^{\prime}<\ldots<z_{n+1}^{\prime}

the roots of polynomialsf,gf^{\prime},g^{\prime}. Eitherkkthe index determined such that

x1<x2<<xk1<has<b<xk<xk+1<<xnx_{1}<x_{2}<\ldots<x_{k-1}<a<b<x_{k}<x_{k+1}<\ldots<x_{n}

if1<k<n+11<k<n+1and let's askk=1k=1if all the rootsxix_{i}are to the right ofbbAndk=n+1k=n+1if all the rootsxix_{i}are to the left ofhasaThe natural numberkkis well determined and takes the values1,2,,n+11,2,\ldots,n+1. SOzkz_{k}is the root oftt^{\prime}which is betweenhasaAndbbAndzkz_{k}^{\prime}the roots ofgg^{\prime}which is betweenhas+ρa+\rhoAndbρb-\rhoThe other root pairsziz_{i},ziz_{i}^{\prime}are respectively included in the open intervals:

(xi,xi+1), For i=1,2,,k2,\displaystyle\left(x_{i},x_{i+1}\right),\quad\text{ pour }i=1,2,\ldots,k-2,
(xk1,has+b2), For i=k1,\displaystyle\left(x_{k-1},\frac{a+b}{2}\right),\text{ pour }i=k-1,
(has+b2,xk), For i=k+1,\displaystyle\left(\frac{a+b}{2},x_{k}\right),\text{ pour }i=k+1,
(xi2,xi1), For i=k+2,k+3,,n+1.\displaystyle\left(x_{i-2},x_{i-1}\right),\text{ pour }i=k+2,k+3,\ldots,n+1.

In this table we remove the first two rows ifk=1k=1, 1a first line ifk=2k=2, the last line ifk=nk=nand the last two lines ifk=n+1k=n+1Finally, forn=1n=1Andn=2n=2we keep one or both of the second and third lines.

The formulas

{f=(xhas)(xb)h+(2xhasb)hg=(xhasρ)(xb+ρ)h+(2xhasb)h\left\{\begin{array}[]{l}f^{\prime}=(x-a)(x-b)h^{\prime}+(2x-a-b)h\\ g^{\prime}=(x-a-\rho)(x-b+\rho)h^{\prime}+(2x-a-b)h\end{array}\right.

show us, sinceh,hh,h^{\prime}cannot have common roots, thatf,gf^{\prime},g^{\prime}can only havehas+b2\frac{a+b}{2}as a common root and this if and setily ifhh^{\prime}cancels out forx=has+b2x=\frac{a+b}{2}. SOzk=zk=has+b2z_{k}=z_{k}^{\prime}=\frac{a+b}{2}. Ifiki\neq k, We haveziziz_{i}\neq z_{i}^{\prime}and for such ai,zi,zii,z_{i},z_{i}^{\prime}cannot be roots ofhh^{\prime}Moreover, we have the formula

f=gρ(bhasρ)h,f^{\prime}=g^{\prime}-\rho(b-a-\rho)h^{\prime}, (22)

which results from (21).
To further study the pairszi,ziz_{i},z_{i}^{\prime}Foriki\neq kWe will distinguish between two cases:

Case 1. Suppose thatk>1k>1and let's examine the rootszi,ziz_{i},z_{i}^{\prime}Fori<ki<kFrom the second formula (21) it follows, for suchii,

h(zi)h(zi)>0h\left(z_{i}^{\prime}\right)h^{\prime}\left(z_{i}^{\prime}\right)>0 (23)

and from the first formula (21) and from formula (22) it follows that

sgf(xi)=sgh(xi),sgf(zi)=sgh(zL),\operatorname{sg}f^{\prime}\left(x_{i}\right)=\operatorname{sg}h^{\prime}\left(x_{i}\right),\operatorname{sg}f^{\prime}\left(z_{i}^{\prime}\right)=-\operatorname{sg}h^{\prime}\left(z_{l}^{\prime}\right), (24)

using the sg function herexxwhich, by definition, is equal to1,0-1,0respectjvement 1 following thatxxEast<<,==respectively>0>0.

From (23) it follows thatzLz_{l}^{\prime}is in the right-hand neighborhood ofziz_{i}, more precisely in the interval (xi,yix_{i},y_{i}). We then have.

h(xi)h(zi)>0h^{\prime}\left(x_{i}\right)h^{\prime}\left(z_{i}^{\prime}\right)>0

and from (24) it follows that

sgf(xi)f(zi)=1\operatorname{sg}f^{\prime}\left(x_{i}\right)f^{\prime}\left(z_{i}^{\prime}\right)=-1 (26)

which shows us thatff^{\prime}has at least one root in the interval (xi,zix_{i},z_{i}^{\prime}). But, we can only have one root which is none other thanziz_{i}verifying this property. It therefore follows that we have

zi<zi,i=1,2,,k1z_{i}<z_{i}^{\prime},i=1,2,\ldots,k-1 (27)

Ifk=u+1k=u+1, asyny_{n}we can take the improper point\inftyin the preceding considerations.

Case 2. Suppose thatk<n+1k<n+1and let's examine the tacineszi,ziz_{i},z_{i}^{\prime}Fori>ki>kBy proceeding as above, we see that instead of (23) we have, for these values ​​ofii,

h(zi)h(zi)<0h\left(z_{i}\right)\cdot h^{\prime}\left(z_{i}^{\prime}\right)<0 (\prime)

which shows us thatzLz_{l}^{\prime}is in the left neighborhood of the pointxL1x_{l-1}, more precisely that it is in the interval(yi2,xi1)\left(y_{i-2},x_{i-1}\right)Instead of (24), (25) and (26) we have respectively

sgf(xi1)=sgh(xi1),sgf(zi)=sgh(zi),\operatorname{sg}f^{\prime}\left(x_{i-1}\right)=\operatorname{sg}h^{\prime}\left(x_{i-1}\right),\operatorname{sg}f^{\prime}\left(z_{i}^{\prime}\right)=-\operatorname{sg}h^{\prime}\left(z_{i}^{\prime}\right), (\prime)
h(xi1)h(zi)>0h^{\prime}\left(x_{i-1}\right)h^{\prime}\left(z_{i}^{\prime}\right)>0 (\prime)
sgf(xi1)f(zi)=1\operatorname{sg}f^{\prime}\left(x_{i-1}\right)f^{\prime}\left(z_{i}^{\prime}\right)=-1 (\prime)

and we deduce, as above, thatziz_{i}is in the interval (zL,xi1z_{l}^{\prime},x_{i-1}).

So we have,

zi<zi,i=k+1,k+2,,n+1z_{i}^{\prime}<z_{i},i=k+1,k+2,\ldots,n+1 (\prime)

Ifk=1k=1, asy0y_{0}we can take the improper point-\inftyin our considerations.

Inequalities (27), (27') and equality

z1+z2++zn+1=z1+z2++zn+1z_{1}+z_{2}+\ldots+z_{n+1}=z_{1}^{\prime}+z_{2}^{\prime}+\ldots+z_{n+1}^{\prime}

demonstrate Lemma 6. Indeed, by virtue of this equality, the inequalities

z1+z2++zi<z1+z2++zi,i=1,2,,n,z_{1}+z_{2}+\ldots+z_{i}<z_{1}^{\prime}+z_{2}^{\prime}+\ldots+z_{i}^{\prime},i=1,2,\ldots,n,

are equivalent to inequalities

{z1+z2++zi<z1+z2++zi,i=1,2,,k1,zi+1+zi+2++zn+1<zi+1+zi+2++zn+1,i=k,k+1,,n\left\{\begin{array}[]{l}z_{1}+z_{2}+\ldots+z_{i}<z_{1}^{\prime}+z_{2}^{\prime}+\ldots+z_{i}^{\prime},i=1,2,\ldots,k-1,\\ z_{i+1}^{\prime}+z_{i+2}^{\prime}+\ldots+z_{n+1}^{\prime}<z_{i+1}+z_{i+2}+\ldots+z_{n+1},i=k,k+1,\ldots,n\end{array}\right.

Ifk=1k=1we eliminate inequalities (27) and the first inequalities (28) and ifk=n+1k=n+1We eliminate the inequalities (27') and the last inequalities (28). The reasoning remains valid.

Theorem 6 is therefore completely proven.
Note: Case 2 can be deduced from Case 1 if we use the property that when the roots of a polynomial undergo a linear transformation, the roots of the derivative of that polynomial undergo the same transformation for any particular values ​​ofhasaandbb(not only for example forhas=0,b=1a=0,b=1or forhas=1,b=1a=-1,b=1).
19. From previous results we can deduce consequences for the case where we have (12) or (12'), but equality (13) transforms into an inequality.

The relationshipPmcQP\xrightarrow{mc}Qmeans that the roots (1) of the polynomialsP,QP,Qdegreen1n\geqq 1verify the inequalities

x1+x2++xiy1+y2++yi,i=1,2,,nx_{1}+x_{2}+\ldots+x_{i}\leqq y_{1}+y_{2}+\ldots+y_{i},i=1,2,\ldots,n

and the relationshipP memc QP\xrightarrow{\text{ memc }}Qmeans that the roots(1)\left(1^{\prime}\right)polynomialsP,QP,Qof degreen1n\geqq 1verify the inequalities

x1+x2++xi<y1+y2++yi,i=1,2,,n.x_{1}+x_{2}+\ldots+x_{i}<y_{1}+y_{2}+\ldots+y_{i},i=1,2,\ldots,n.

The relationshipmc\xrightarrow{mc}is (reflexive and) transitive and the relationmemc\xrightarrow{memc}is also transitive. OfPcQP\xrightarrow{\mathrm{c}}QrespectivelyPccQP\xrightarrow{\mathrm{cc}}Qit resultsPmcQP\xrightarrow{\mathrm{mc}}QrespectivelyPmcmcQP\xrightarrow{mcmc}QandPmQP\xrightarrow{m}Qit follows thatPmcQP\xrightarrow{mc}Q, etc.

We then have
Consequence 4. IfP,QP,Qare two polynomials of degreen>1n>1, ofPmcQP\xrightarrow{\mathrm{mc}}Qit follows thatPmcQP^{\prime}\xrightarrow{\mathrm{mc}}Q^{\prime}.

Indeed, ifRRis a polynomial of degreennhaving as its rootsx1,x2,,xn1,y1+y2++yn(x1+x2++xn1)x_{1},x_{2},\ldots,x_{n-1},y_{1}+y_{2}+\ldots+y_{n}-\left(x_{1}+x_{2}+\ldots+x_{n-1}\right), We havePcR,RmQP\xrightarrow{\mathrm{c}}R,R\xrightarrow{\mathrm{~m}}QSo we havePcR,RQP^{\prime}\xrightarrow{\mathrm{c}}R^{\prime},R^{\prime}\rightarrow Q^{\prime}As a result,PmcRP^{\prime}\xrightarrow{\mathrm{mc}}R^{\prime},RmcQR^{\prime}\xrightarrow{mc}Q^{\prime}, hencePmcQP^{\prime}\xrightarrow{mc}Q^{\prime}, which demonstrates consequence 4.

We also have
Consequence 5. IfP,QP,Qare two polynomials of degreen>1n>1, ofP mcmc QP\xrightarrow{\text{ mcmc }}Qit follows thatP mcmc QP^{\prime}\xrightarrow{\text{ mcmc }}Q^{\prime}.

EitherRRthe previous polynomial. In this case, we havePRP^{\prime}\rightarrow R^{\prime}, without equality (13) being verified, andRmnQR^{\prime}\xrightarrow{mn}Q^{\prime}It is easy to see that the relationshipP memc QP^{\prime}\xrightarrow{\text{ memc }}Q^{\prime}The result is...

Consequence 4 can be deduced from consequence 5. Let (14) denote the roots of the polynomialPPdegreennand eitherPQP_{Q}a polynomial of degreennhaving as its rootsxi(ni)px_{i}-(n-i)p, Orppis a positive number. The polynomialPQP_{Q}has all its simple roots and ifρ0\rho\rightarrow 0the roots ofPQP_{Q}respectively those ofPP_{\circ}^{\prime}tend towards the roots ofPPrespectively towards those ofPP^{\prime}A simple calculation, similar to that in no. 13, shows us that ifPmcQP\xrightarrow{mc}Qwe also haveP2ϱmcmcQQP_{2\varrho}\xrightarrow{mcmc}Q_{Q}Assuming consequence 5 is proven, we haveP2ϱ mcmc QϱP_{2\varrho}^{\prime}\xrightarrow{\text{ mcmc }}Q_{\varrho}, hence, by doingρ0\rho\rightarrow 0, we deducePmcQP^{\prime}\xrightarrow{\mathrm{mc}}Q^{\prime}, which demonstrates consequence 4.

Finally, consequence 5 can also be demonstrated by taking into account Lemma 1, Theorem 6, and by increasing the last rootxnx_{n}of the polynomialPPWe ask the reader to perform this demonstration.

BIBLIOGRAPHICAL

[1] Hardy GH, Littlewood JE, Pó1y a G., Inequalities, 1934.
[2] Markov WA, Über Polynome die in einen gegebenen Intervalle mőglichst wenig von Null abweichen. Mathematische Annalen, 77, 213-258 (1916).
[3] Montel P., On rational fractions with interleaved terms, Mathematica, 5, 110-129 (1931). Received on 11. VI. 1960.

9 - Mathematica

1961

Related Posts