Chapter 5.2

The Geometry of the Complex Plane and the Zeros of Polynomials

[56/100%]
Show
LaTeX
Problem 5.101

Let z1,z2,…,znz_1, z_2, \ldots , z_n be arbitrary points lying in the finite part of the complex plane and let m1,m2,…,mnm_1, m_2, \ldots , m_n be non-negative masses whose sum is 11 which are placed at the points z1,z2,…,znz_1, z_2, \ldots , z_n respectively. In any linear transformation of the complex plane that carries the point at infinity into itself, the center of gravity ζ\zeta of such a mass-distribution is mapped by the same transformation, i.e. if we again place the point mass mνm_{\nu } at the image-point zν′z_{\nu }' of zνz_{\nu } (1≤ν≤n1 \leq \nu \leq n), then the center of gravity ζ′\zeta ' of this new mass distribution is identical with the image of the point ζ\zeta.

?
Problem 5.102

In the sequel (102--156) we mean by a point'' of the complex plane an arbitrary finite point or the point at infinity, by a circle'' the circumference of a circle or a straight line---the latter passes through the point at infinity---and by a ``circular domain'' a domain whose boundary is a circle. A circular domain is either the closed interior or the closed exterior of a circle or a closed half-plane, according as it does not contain the point at infinity or contains it in its interior or on its boundary.

Let the points z1,z2,…,zn,zz_1, z_2, \ldots , z_n, z and the masses m1,m2,…,mnm_1, m_2, \ldots , m_n be given. Let zz be different from all the points z1,z2,…,znz_1, z_2, \ldots , z_n while of the latter some may coincide. Let m1≥0,m2≥0,…,mn≥0,m1+m2+⋯+mn=1m_1 \geq 0, m_2 \geq 0, \ldots , m_n \geq 0, m_1+m_2+\cdots +m_n=1.

Find a point ζ\zeta situated in such a manner that, in a linear transformation of the plane that maps the n+2n+2 points

z1,z2,z3,…,zn;z;ζ z_1, z_2, z_3, \ldots , z_n; \quad z; \quad \zeta

into the points

z1′,z2′,z3′,…,zn′;∞;ζ′ z_1', z_2', z_3', \ldots , z_n'; \quad \infty ; \quad \zeta '

respectively, ζ′\zeta ' becomes the center of gravity of the mass-distribution determined by the masses mνm_{\nu } placed at the points zν′z_{\nu }' (ν=1,2,3,…,n\nu =1,2,3,\ldots ,n).

There are infinitely many linear transformations that map zz into ∞\infty; the point ζ\zeta however is independent of the choice of the underlying transformation.

The point ζ=ζz\zeta =\zeta_z, defined above, which is determined uniquely by the system of points z1,z2,…,znz_1, z_2, \ldots , z_n, by the masses m1,m2,…,mnm_1, m_2, \ldots , m_n, m1+m2+⋯+mn=1m_1+m_2+\cdots +m_n=1, placed there, and by the ``point of reference'' zz which must be distinct from all the points zνz_{\nu }, is called the center of gravity of the mass distribution m1,m2,…,mnm_1, m_2, \ldots , m_n with respect to zz. If zz is the point at infinity, z=∞z=\infty, then ζz=ζ∞\zeta_z=\zeta_{\infty } is the ordinary center of gravity.

?
Problem 5.103

Consider all possible mass distributions with total mass 11 over the fixed points z1,z2,…,znz_1, z_2, \ldots , z_n and let zz be a point distinct from z1,z2,…,znz_1, z_2, \ldots , z_n. The centers of gravity ζz\zeta_z of all mass distributions of this kind with respect to zz cover a ``circular-arc polygon'' Cz\mathfrak {C}_z.

?
Problem 5.104

Let Cz\mathfrak {C}_z be the circular-arc polygon defined in 103 and w1,w2w_1, w_2 two points of Cz\mathfrak {C}_z. Consider the circle through w1,w2w_1, w_2 and zz. The arc of this circle with end-points w1,w2w_1, w_2, that does not contain zz, is contained in Cz\mathfrak {C}_z, in other words Cz\mathfrak {C}_z is circularly-convex with respect to zz. (We call Cz\mathfrak {C}_z the smallest circularly-convex domain with respect to zz that contains the points z1,z2,…,znz_1, z_2, \ldots , z_n.)

?
Problem 5.105

If the points z1,z2,…,znz_1, z_2, \ldots , z_n all belong to a circular domain CC and if the point zz lies outside CC, then the circular-arc polygon Cz\mathfrak {C}_z [103] lies completely inside CC.

?
Problem 5.106

Let z1,z2,…,znz_1, z_2, \ldots , z_n be arbitrary points and zz a point distinct from these. Let the same mass 1/n1/n be placed at each of the points zνz_{\nu }. In the sequel we consider only the center of gravity ζ=ζz\zeta =\zeta_z of this special mass distribution and call it for brevity the center of gravity of the points z1,z2,…,znz_1, z_2, \ldots , z_n with respect to zz.

Let ζz\zeta_z be the center of gravity of z1,z2,…,znz_1, z_2, \ldots , z_n with respect to zz. Every circle through zz and ζz\zeta_z separates the points z1,z2,…,znz_1, z_2, \ldots , z_n, i.e. either both circular domains bounded by the circle contain some of the points z1,z2,…,znz_1, z_2, \ldots , z_n in their interior or all the points z1,z2,…,znz_1, z_2, \ldots , z_n lie on the circle.

?
Problem 5.107

If all the points z1,z2,…,znz_1, z_2, \ldots , z_n belong to a circular domain CC, then the points zz and ζz\zeta_z cannot both lie outside CC. If one of them, e.g. zz, lies outside CC, then the other, ζz\zeta_z, lies in the interior of CC, except in the case where all the points z1,z2,…,znz_1, z_2, \ldots , z_n coincide at a boundary point of CC and ζz\zeta_z then also lies there.

?
Problem 5.108

Let the fixed points w1,w2,…,wk,zw_1, w_2, \ldots , w_k, z be pair-wise distinct. Let the complex numbers z1,z2,…,znz_1, z_2, \ldots , z_n take only the kk different values w1,w2,…,wkw_1, w_2, \ldots , w_k. The centers of gravity with respect to zz of all systems z1,z2,…,znz_1, z_2, \ldots , z_n of this kind lie everywhere dense in the smallest domain, circularly convex with respect to zz, that contains the points w1,w2,…,wkw_1, w_2, \ldots , w_k.

?
Problem 5.109

The points z1,z2,…,znz_1, z_2, \ldots , z_n are fixed and the variable point zz converges to z1z_1. To what limiting position does the center of gravity ζz\zeta_z of the points z1,z2,…,znz_1, z_2, \ldots , z_n with respect to zz tend?

?
Problem 5.110

The center of gravity ζz\zeta_z of the finite fixed points z1,z2,…,znz_1, z_2, \ldots , z_n with respect to the variable point zz may for sufficiently large zz be expanded in decreasing powers of zz. Calculate the first two terms of the expansion.

?
Problem 5.111

In the sequel (111--156) we shall mean by a rational integral function (polynomial) of degree nn a function f(z)f(z) not vanishing identically of the form

f(z)=a0+(n1)a1z+(n2)a2z2+⋯+(nn−1)an−1zn−1+anzn. f(z) = a_0 + \binom {n}{1}a_1z + \binom {n}{2}a_2z^2 + \cdots + \binom {n}{n-1}a_{n-1}z^{n-1} + a_nz^n.

Here ana_n is not necessarily different from 00. In the special case when an≠0a_n \neq 0, we say that nn is the precise degree of f(z)f(z). If an=an−1=⋯=an−k+1=0a_n=a_{n-1}=\cdots =a_{n-k+1}=0, an−k≠0a_{n-k} \neq 0, we shall say that f(z)f(z) has the kk-fold zero z=∞z=\infty. With this interpretation every rational integral function of degree nn has a total of nn zeros, of which some may possibly lie at infinity. Thus there is determined by f(z)f(z) a system of points z1,z2,…,znz_1, z_2, \ldots , z_n, in the complex plane, the zeros of f(z)f(z). Conversely, to every system of nn points there belongs a polynomial, whose coefficients are uniquely determined up to a constant of proportion. We shall often disregard this unimportant constant factor and speak of the polynomial, whose zeros are z1,z2,…,znz_1, z_2, \ldots , z_n.

By the center of gravity ζ\zeta of a polynomial we mean the center of gravity of its zeros. Express the center of gravity of f(z)f(z) with respect to a variable point zz: (1) in terms of f(z)f(z) and f′(z)f'(z), (2) in terms of the coefficients of f(z)f(z). Consider the special cases z=∞z=\infty and f(z)=znf(z)=z^n.

?
Problem 5.112

Let ζ\zeta be the center of gravity of f(z)f(z) with respect to zz. A necessary and sufficient condition that f(z)f(z) has only real zeros is that the imaginary parts of zz and ζ\zeta have opposite signs if they do not both vanish however zz may vary. (z=∞z=\infty counts as a real zero.)

?
Problem 5.113

With respect to which points zz of the plane does the center of gravity of f(z)f(z) lie at infinity? What do the statements in 106 and 107 say about these points?

?
Problem 5.114

Let f(z)f(z) be a polynomial of precise degree nn, CC a circular domain that contains the zeros of f(z)f(z) and c≠0c \neq 0. The zeros of the derivative of the transcendental function e−z/cf(z)e^{-z/c}f(z) lie either in CC or in C+ncC+nc, i.e. in the circular domain that is obtained from CC by parallel displacement by the vector ncnc.

?
Remark.
?

Cf. Problem 33 of Part Three (Gauss's theorem in the circular-domain formulation used in 113); consider the center of gravity of f(z)f(z) with respect to one of the zeros in question.

Problem 5.115

Let z1z_1 be a zero of the polynomial f(z)f(z) of degree nn, xx finite and x≠z1x \neq z_1, and f′(x)=0f'(x)=0. Then f(z)f(z) has at least one zero in every circle that passes through both the points

xandx−(n−1)(z1−x). x \quad \text{and} \quad x-(n-1)(z_1-x).
?
Problem 5.116

Let f(z)f(z) be a polynomial of degree nn, whose zeros are all of absolute value ≥1\geq 1. Let α1\alpha_1 and α2\alpha_2 be arbitrary positive numbers, such that nα1/α2≠1n\alpha_1/\alpha_2 \neq 1. Then all the zeros of α1zf′(z)−α2f(z)\alpha_1zf'(z)-\alpha_2f(z) are in absolute value

≥min⁡(1,∣1−nα1α2∣−1). \geq \min \left(1, \left|1-n\frac{\alpha _1}{\alpha _2}\right|^{-1}\right).
?
Problem 5.117

Let f(z)f(z) be a polynomial of degree nn, all of whose zeros lie in the annulus r≤∣z∣≤Rr \leq \left|z\right| \leq R. Let α1\alpha_1 and α2\alpha_2 be arbitrarily positive numbers such that nα1/α2≠1n\alpha_1/\alpha_2 \neq 1. Then all the zeros of α1zf′(z)−α2f(z)\alpha_1zf'(z)-\alpha_2f(z) lie in the annulus

rmin⁡(1,∣1−nα1α2∣−1)≤∣z∣≤Rmax⁡(1,∣1−nα1α2∣−1). r\min \left(1, \left|1-n\frac{\alpha _1}{\alpha _2}\right|^{-1}\right) \leq \left|z\right| \leq R\max \left(1, \left|1-n\frac{\alpha _1}{\alpha _2}\right|^{-1}\right).
?
Problem 5.118

If z1z_1 is a finite zero of the polynomial f(z)f(z) of degree nn, then the center of gravity with respect to z1z_1 of the remaining n−1n-1 zeros of f(z)f(z) is

=z1−2(n−1)f′(z1)f′′(z1). = z_1 - \frac{2(n-1)f'(z_1)}{f''(z_1)}.

How is this formula to be modified if z1z_1 is at infinity?

?
Problem 5.119

The nnth Hermite polynomial satisfies the differential equation

f′′(z)−zf′(z)+nf(z)=0. f''(z)-zf'(z)+nf(z)=0.

(This is the differential equation satisfied by the Hermite polynomial as defined via e−z2/2Hn(z)=1n!dndzne−z2/2e^{-z^2/2}H_n(z) = \tfrac {1}{n!}\tfrac {d^n}{dz^n}e^{-z^2/2} in problem 100(g) of Part Six's solution.) Show by means of 118 that the Hermite polynomials have only real zeros (problem 100(i) of Part Six's solution). [Consider the assumed complex zero with largest imaginary part.]

?
Problem 5.120

The polynomial of degree nn that satisfies the differential equation

(1−z2)f′′(z)−2zf′(z)+n(n+1)f(z)=0 (1-z^2)f''(z)-2zf'(z)+n(n+1)f(z)=0

(the nnth Legendre polynomial, cf. problem 90 of Part Six), has only real zeros (problem 97 of Part Six).

?
Problem 5.121

Gauss's theorem (the circular-domain formulation given in the introduction to problem 31 of Part Three) is equivalent to the following: If all the zeros of a polynomial f(z)f(z) lie in a circular disc CC, then the zeros of its derivative f′(z)f'(z) also lie in CC [113]. Decide whether the following theorem is true or false: If all the zeros of a polynomial f(z)f(z) lie in two circular discs C1,C2C_1, C_2, then the zeros of its derivative f′(z)f'(z) also lie in C1C_1 or C2C_2.

?
Problem 5.122

Let C1C_1 and C2C_2 be two circular discs or, more generally, two arbitrary circular domains. Denoting by n1n_1 and n2n_2 fixed positive numbers, we call the set of all points

z=n1z2+n2z1n1+n2, z = \frac{n_1z_2+n_2z_1}{n_1+n_2},

where z1z_1 and z2z_2 independently run through C1C_1 and C2C_2 respectively, a mean domain of C1C_1 and C2C_2, and use for it the following notation

C=n1C2+n2C1n1+n2. C = \frac{n_1C_2+n_2C_1}{n_1+n_2}.

CC is completely determined by C1,C2C_1, C_2 and n1,n2n_1, n_2. [Cf. H. Minkowski: Werke, Vol. 2, p. 176, Leipzig: B. G. Teubner 1911.]

Let C1C_1 and C2C_2 be two circular discs whose centers are z1(0),z2(0)z_1^{(0)}, z_2^{(0)} and radii r1,r2r_1, r_2 respectively. Show that then

C=n1C2+n2C1n1+n2 C = \frac{n_1C_2+n_2C_1}{n_1+n_2}

is again a circular disc whose center z(0)z^{(0)} and radius rr are given by the following equations:

z(0)=n1z2(0)+n2z1(0)n1+n2,r=n1r2+n2r1n1+n2. z^{(0)} = \frac{n_1z_2^{(0)}+n_2z_1^{(0)}}{n_1+n_2}, \qquad r = \frac{n_1r_2+n_2r_1}{n_1+n_2}.

CC is homothetic to C1C_1 and C2C_2 with respect to the same center of similitude. [H. S. M. Coxeter: Introduction to Geometry. New York: John Wiley Sons 1961, pp. 68--70.]

?
Problem 5.123

Let C1C_1 and C2C_2 be two half-planes with parallel boundary lines, one of which contains the other. The mean domain C=n1C2+n2C1/(n1+n2)C=n_1C_2+n_2C_1/(n_1+n_2) is then again a half-plane, whose boundary line runs parallel to those of C1C_1 and C2C_2 and divides the distance between them in the ratio n1:n2n_1:n_2.

?
Problem 5.124

Let f(z)=f1(z)f2(z)f(z)=f_1(z)f_2(z), and f1(z)f_1(z) be a polynomial of degree n1n_1, f2(z)f_2(z) a polynomial of degree n2n_2. Let all the zeros of f1(z)f_1(z) lie in the circular domain C1C_1 and all the zeros of f2(z)f_2(z) in the circular domain C2C_2. Then the zeros of the derivative f′(z)f'(z) lie either in C1C_1 or in C2C_2 or in the mean domain

C=n1C2+n2C1n1+n2. C = \frac{n_1C_2+n_2C_1}{n_1+n_2}.

(By degree we mean the precise degree.)

?
Problem 5.125

Let f(z)f(z) be a rational (non-integral) function, whose numerator and denominator have no common factors. Let it have n1n_1 zeros (poles) which lie in the circular domain C1C_1 and n2n_2 poles (zeros) which lie in the circular domain C2C_2. (Only finite zeros and poles are considered.) Let n1≥n2n_1 \geq n_2. Then the zeros of f′(z)f'(z) lie either in C1C_1 or in C2C_2 or in the domain CC defined by the numbers

z=n1z2−n2z1n1−n2, z = \frac{n_1z_2-n_2z_1}{n_1-n_2},

where z1z_1 and z2z_2 independently run through C1C_1 and C2C_2, for which we use the notation

C=n1C2−n2C1n1−n2. C = \frac{n_1C_2-n_2C_1}{n_1-n_2}.

If n1=n2n_1=n_2 and the circular domains C1C_1 and C2C_2 have no common points, then all the zeros of f′(z)f'(z) lie either in C1C_1 or in C2C_2.

?
Problem 5.126

Let f(z)f(z) be a polynomial, and let the equation

f(z)=a f(z)=a

have all its roots in a convex domain (oval) O1\mathfrak {O}_1. Then all the numbers aa for which this is the case also lie in a convex domain O2\mathfrak {O}_2.

?
Problem 5.127

Let f(z)f(z) be a polynomial and let its aa-points (the zeros of f(z)−af(z)-a) lie in a circular domain C1C_1 and its bb-points in a circular domain C2C_2. If n1,n2n_1, n_2 are positive numbers and we set c=(n1b+n2a)/(n1+n2)c=(n_1b+n_2a)/(n_1+n_2), those cc-points of f(z)f(z) that lie neither in C1C_1 nor in C2C_2, lie in the mean domain

C=n1C2+n2C1n1+n2 C = \frac{n_1C_2+n_2C_1}{n_1+n_2}

[We may take n1,n2n_1, n_2 to be integers.]

?
Problem 5.128

By the derivative of a polynomial f(z)f(z) of degree nn with respect to the point ζ\zeta, denoted by Aζf(z)A_{\zeta }f(z), we mean the ordinary derivative if ζ=∞\zeta =\infty, and

Aζf(z)=(ζ−z)f′(z)+nf(z), A_{\zeta }f(z) = (\zeta -z)f'(z)+nf(z),

if ζ\zeta is finite. Aζf(z)A_{\zeta }f(z) is a polynomial of degree n−1n-1. (When using homogeneous coordinates this expression is called the first polar, or, by older authors, emanant; cf. Laguerre, Oeuvres, Vol. 1, p. 48. Paris: Gauthier-Villars, 1898.)

Let the n+1n+1 coefficients of f(z)f(z) (in the notation of §2 preceding 111) be a0,a1,a2,…,ana_0, a_1, a_2, \ldots , a_n. What are the nn coefficients of [Aζf(z)]/n[A_{\zeta }f(z)]/n?

?
Problem 5.129

Let ζ\zeta be arbitrary. Show that the operation Aζf(z)A_{\zeta }f(z) is distributive: If we denote by f1(z)f_1(z) and f2(z)f_2(z) two polynomials of degree nn, and by c1c_1 and c2c_2 arbitrary constants, then we have

Aζ[c1f1(z)+c2f2(z)]=c1Aζf1(z)+c2Aζf2(z). A_{\zeta }[c_1f_1(z)+c_2f_2(z)] = c_1A_{\zeta }f_1(z)+c_2A_{\zeta }f_2(z).
?
Problem 5.130

Let ζ\zeta be arbitrary and f(z)f(z) a polynomial of degree nn, f(z)=g(z)h(z)f(z)=g(z)h(z) a factorization of f(z)f(z) into the two factors g(z)g(z) and h(z)h(z) of degree kk and ll respectively, k+l=nk+l=n. Show that

Aζf(z)=g(z)Aζh(z)+h(z)Aζg(z). A_{\zeta }f(z) = g(z)A_{\zeta }h(z)+h(z)A_{\zeta }g(z).
?
Problem 5.131

Let f(z)f(z) be a polynomial of degree nn, ζ1\zeta_1 and ζ2\zeta_2 two arbitrary constants. The operations Aζ1f(z)A_{\zeta_1}f(z) and Aζ2f(z)A_{\zeta_2}f(z) are commutative, that is to say

Aζ1[Aζ2f(z)]=Aζ2[Aζ1f(z)]=Aζ1Aζ2f(z). A_{\zeta _1}[A_{\zeta _2}f(z)] = A_{\zeta _2}[A_{\zeta _1}f(z)] = A_{\zeta _1}A_{\zeta _2}f(z).
?
Problem 5.132

Show that Aζf(z)A_{\zeta }f(z) vanishes identically if and only if all the zeros of f(z)f(z) coincide with ζ\zeta.

?
Problem 5.133

By the derived system of the nn points z1,z2,…,znz_1, z_2, \ldots , z_n with respect to an (n+1)(n+1)th point ζ\zeta we mean the n−1n-1 zeros z1′,z2′,…,zn−1′z_1', z_2', \ldots , z_{n-1}' of Aζf(z)A_{\zeta }f(z), where f(z)f(z) is the polynomial of degree nn whose zeros are z1,z2,…,znz_1, z_2, \ldots , z_n, all the zeros being correctly counted (§2, preceding 111). The derived system is completely determined, except in the single exceptional case where the n+1n+1 points z1,z2,…,zn,ζz_1, z_2, \ldots , z_n, \zeta all coincide (cf. 132).

If ζ\zeta is finite, then to the derived system with respect to ζ\zeta in general there belong four different kinds of points:

(1) The finite points distinct from z1,z2,…,zn,ζz_1, z_2, \ldots , z_n, \zeta, with respect to which the center of gravity of z1,z2,…,znz_1, z_2, \ldots , z_n is precisely ζ\zeta.

(2) The multiple zeros of f(z)f(z), distinct from ζ\zeta, each with multiplicity reduced by 11.

(3) ζ\zeta itself only if it is a zero of f(z)f(z) and in that case with the same multiplicity.

(4) The point ∞\infty if it is at least a double zero of f(z)f(z); moreover also if it is not a zero of f(z)f(z) but if ζ\zeta is the ordinary center of gravity of f(z)f(z), i.e. the one constructed with respect to ∞\infty.

?
Problem 5.134

A circular domain, whose boundary contains ζ\zeta as well as a point of the derived system with respect to ζ\zeta, itself contains points of the original system.

?
Problem 5.135

If a circular domain contains the points z1,z2,…,znz_1, z_2, \ldots , z_n and does not contain the point ζ\zeta, then it contains also the derived system of z1,z2,…,znz_1, z_2, \ldots , z_n with respect to ζ\zeta.

?
Problem 5.136

The derived system of the points z1,z2,…,znz_1, z_2, \ldots , z_n lies in the smallest circular-arc polygon, circularly convex with respect to ζ\zeta, that contains z1,z2,…,znz_1, z_2, \ldots , z_n. (Generalization of Problem 31 of Part Three.)

?
Problem 5.137

If f(z)f(z) is of degree nn, then (Aζ)n−1f(z)(A_{\zeta })^{n-1}f(z) is of the first degree. Calculate the only zero of (Aζ)n−1f(z)(A_{\zeta })^{n-1}f(z)!

?
Problem 5.138

Let

f(z)=a0+(n1)a1z+(n2)a2z2+⋯+(nn−1)an−1zn−1+anzn f(z) = a_0+\binom {n}{1}a_1z+\binom {n}{2}a_2z^2+\cdots +\binom {n}{n-1}a_{n-1}z^{n-1}+a_nz^n

be an arbitrary polynomial of degree nn and let ζ1,ζ2,…,ζn\zeta_1, \zeta_2, \ldots , \zeta_n be arbitrary points in the complex plane. The expression

A(ζ1,ζ2,…,ζn)f(z)=1n!Aζ1Aζ2…Aζnf(z) A(\zeta _1, \zeta _2, \ldots , \zeta _n)f(z) = \frac{1}{n!}A_{\zeta _1}A_{\zeta _2}\ldots A_{\zeta _n}f(z)

is a symmetric multilinear function of the variables ζ1,ζ2,…,ζn\zeta_1, \zeta_2, \ldots , \zeta_n, which we can write as follows:

A(ζ1,ζ2,…,ζn)f(z)=a0Σ0+a1Σ1+a2Σ2+⋯+an−1Σn−1+anΣn. A(\zeta _1, \zeta _2, \ldots , \zeta _n)f(z) = a_0\Sigma _0+a_1\Sigma _1+a_2\Sigma _2+\cdots +a_{n-1}\Sigma _{n-1}+a_n\Sigma _n.

Here Σ0,Σ1,Σ2,…,Σn−1,Σn\Sigma_0, \Sigma_1, \Sigma_2, \ldots , \Sigma_{n-1}, \Sigma_n are the elementary symmetric functions of ζ1,ζ2,…,ζn\zeta_1, \zeta_2, \ldots , \zeta_n:

Σ0=1,Σ1=ζ1+ζ2+⋯+ζn, \Sigma _0=1, \qquad \Sigma _1=\zeta _1+\zeta _2+\cdots +\zeta _n, Σ2=ζ1ζ2+ζ1ζ3+⋯+ζn−1ζn,…,Σn=ζ1ζ2…ζn. \Sigma _2=\zeta _1\zeta _2+\zeta _1\zeta _3+\cdots +\zeta _{n-1}\zeta _n, \qquad \ldots , \qquad \Sigma _n=\zeta _1\zeta _2\ldots \zeta _n.

If the last kk of the numbers ζ1,ζ2,…,ζn\zeta_1, \zeta_2, \ldots , \zeta_n (and only these) lie at infinity, then we set Σ0=Σ1=⋯=Σk−1=0\Sigma_0=\Sigma_1=\cdots =\Sigma_{k-1}=0, Σk=1\Sigma_k=1, Σk+1=ζ1+ζ2+⋯+ζn−k\Sigma_{k+1}=\zeta_1+\zeta_2+\cdots +\zeta_{n-k}, Σk+2=ζ1ζ2+ζ1ζ3+⋯+ζn−k−1ζn−k,…,Σn=ζ1ζ2…ζn−k\Sigma_{k+2}=\zeta_1\zeta_2+\zeta_1\zeta_3+\cdots +\zeta_{n-k-1}\zeta_{n-k}, \ldots , \Sigma_n=\zeta_1\zeta_2\ldots \zeta_{n-k}.

?
Problem 5.139

Let

f(z)=a0+(n1)a1z+(n2)a2z2+⋯+(nn−1)an−1zn−1+anzn f(z) = a_0+\binom {n}{1}a_1z+\binom {n}{2}a_2z^2+\cdots +\binom {n}{n-1}a_{n-1}z^{n-1}+a_nz^n

be an arbitrary polynomial of degree nn with zeros z1,z2,…,znz_1, z_2, \ldots , z_n and

g(z)=b0+(n1)b1z+(n2)b2z2+⋯+(nn−1)bn−1zn−1+bnzn g(z) = b_0+\binom {n}{1}b_1z+\binom {n}{2}b_2z^2+\cdots +\binom {n}{n-1}b_{n-1}z^{n-1}+b_nz^n

an arbitrary polynomial of degree nn with zeros ζ1,ζ2,…,ζn\zeta_1, \zeta_2, \ldots , \zeta_n. Calculate A(ζ1,ζ2,…,ζn)f(z)A(\zeta_1, \zeta_2, \ldots , \zeta_n)f(z) and A(z1,z2,…,zn)g(z)A(z_1, z_2, \ldots , z_n)g(z).

If the two polynomials are related in such a manner that

A(ζ1,ζ2,…,ζn)f(z)=0 A(\zeta _1, \zeta _2, \ldots , \zeta _n)f(z) = 0

or, equivalently [solution 139], A(z1,z2,…,zn)g(z)=0A(z_1, z_2, \ldots , z_n)g(z)=0, then the two polynomials f(z)f(z) and g(z)g(z) are termed apolar. The term suggests the vanishing of the nnth polar Aζ1Aζ2…Aζnf(z)A_{\zeta_1}A_{\zeta_2}\ldots A_{\zeta_n}f(z). The condition of apolarity is

a0bn−(n1)a1bn−1+(n2)a2bn−2−⋯+(−1)n−1(nn−1)an−1b1+(−1)nanb0=0. a_0b_n-\binom {n}{1}a_1b_{n-1}+\binom {n}{2}a_2b_{n-2}-\cdots +(-1)^{n-1}\binom {n}{n-1}a_{n-1}b_1+(-1)^na_nb_0=0.

[139.] The two systems z1,z2,…,znz_1, z_2, \ldots , z_n and ζ1,ζ2,…,ζn\zeta_1, \zeta_2, \ldots , \zeta_n are also termed mutually apolar.

?
Problem 5.140

What is the geometrical significance of the apolarity of z1,ζ1z_1, \zeta_1, and that of z1,z2z_1, z_2 and ζ1,ζ2\zeta_1, \zeta_2?

?
Problem 5.141

Which are the systems ζ1,ζ2,…,ζn\zeta_1, \zeta_2, \ldots , \zeta_n, that are apolar to the system consisting of the roots of the equation zn−1=0z^n-1=0?

?
Problem 5.142

Let z1,z2,…,znz_1, z_2, \ldots , z_n be an arbitrary system of points. There are nn systems ζ,ζ,…,ζ\zeta ,\zeta ,\ldots ,\zeta of coincident points that are apolar to the given system. Which are they?

?
Problem 5.143

The polynomial of degree nn

f(z)=1−z+czn f(z) = 1-z+cz^n

is apolar to any system of numbers whose sum is nn and product is 00.

?
Problem 5.144

Let f(z)f(z) be an arbitrary polynomial of degree nn, all of whose zeros lie in the circular domain CC. If ζ1,ζ2,…,ζk\zeta_1, \zeta_2, \ldots , \zeta_k, k<nk<n, are arbitrary points lying outside CC, then all the zeros of the polynomial of degree n−kn-k, Aζ1Aζ2…Aζkf(z)A_{\zeta_1}A_{\zeta_2}\ldots A_{\zeta_k}f(z) lie in CC.

?
Problem 5.145

Let the two polynomials f(z)f(z) and g(z)g(z) of degree nn be apolar to each other. Every circular domain that contains all the zeros of one polynomial also contains at least one zero of the other.

?
Problem 5.146

Let the two polynomials f(z)f(z) and g(z)g(z) of degree nn be apolar to each other. The two smallest convex polygons that contain all the zeros of f(z)f(z) and g(z)g(z) respectively have at least one common point. More generally, the two smallest circular-arc polygons that are circularly convex with respect to the same point and that contain all the zeros of f(z)f(z) and g(z)g(z) respectively, also have at least one common point.

?
Problem 5.147

The polynomial

1−z+czn 1-z+cz^n

always has one zero in the circle ∣z∣≤2\left|z\right| \leq 2.

?
Problem 5.148

The polynomial

1−z+czn 1-z+cz^n

always has one zero in the circle ∣z−1∣≤1\left|z-1\right| \leq 1.

?
Problem 5.149

The polynomial of k+1k+1 terms

1−z+c2zν2+c3zν3+⋯+ckzνk,1=ν1<ν2<ν3<⋯<νk 1-z+c_2z^{\nu _2}+c_3z^{\nu _3}+\cdots +c_kz^{\nu _k}, \qquad 1=\nu _1<\nu _2<\nu _3<\cdots <\nu _k

always has one zero in the circle

∣z∣≤[(1−1ν2)(1−1ν3)⋯(1−1νk)]−1, \left|z\right| \leq \left[\left(1-\frac{1}{\nu _2}\right)\left(1-\frac{1}{\nu _3}\right)\cdots \left(1-\frac{1}{\nu _k}\right)\right]^{-1},

and hence in the circle

∣z∣≤k. \left|z\right| \leq k.

(Generalization of 147.)

?
Problem 5.150

If a polynomial of degree nn assumes the same value at two points aa and bb, a≠ba \neq b, then its derivative has at least one zero in the circular disc with center at the midpoint of the line segment abab and radius (∣a−b∣/2)cot⁡(π/n)(\left|a-b\right|/2)\cot (\pi /n). (Analogue of Rolle's theorem in the complex domain.)

?
Problem 5.151

Let

f(z)=a0+(n1)a1z+(n2)a2z2+⋯+(nn−1)an−1zn−1+anzn f(z) = a_0+\binom {n}{1}a_1z+\binom {n}{2}a_2z^2+\cdots +\binom {n}{n-1}a_{n-1}z^{n-1}+a_nz^n

be a polynomial of degree nn all of whose zeros lie in a circular domain CC. Also let

g(z)=b0+(n1)b1z+(n2)b2z2+⋯+(nn−1)bn−1zn−1+bnzn g(z) = b_0+\binom {n}{1}b_1z+\binom {n}{2}b_2z^2+\cdots +\binom {n}{n-1}b_{n-1}z^{n-1}+b_nz^n

be a polynomial of degree nn with zeros β1,β2,…,βn\beta_1, \beta_2, \ldots , \beta_n. Every zero γ\gamma of the polynomial

h(z)=a0b0+(n1)a1b1z+(n2)a2b2z2+⋯+(nn−1)an−1bn−1zn−1+anbnzn h(z) = a_0b_0+\binom {n}{1}a_1b_1z+\binom {n}{2}a_2b_2z^2+\cdots +\binom {n}{n-1}a_{n-1}b_{n-1}z^{n-1}+a_nb_nz^n

formed by ``composition'' of f(z)f(z) and g(z)g(z) then has the form

γ=−βνk, \gamma = -\beta _{\nu }k,

where ν\nu denotes a suitably chosen index and kk a suitably chosen point in CC. (Here we must set ∞⋅∞=∞\infty \cdot \infty = \infty and 0⋅∞=0 \cdot \infty = indeterminate.)

?
Problem 5.152

Let the zeros of the polynomials f(z)f(z) and g(z)g(z) of degree nn all lie in the unit circle ∣z∣≤1\left|z\right| \leq 1, and the zeros of at least one of the two polynomials in fact in ∣z∣<1\left|z\right|<1. Then also all the zeros of the polynomial of degree nn formed from f(z)f(z) and g(z)g(z) by composition [151] lie in the interior of the unit circle ∣z∣<1\left|z\right|<1.

?
Problem 5.153

Let the polynomial f(z)f(z) of degree nn have all its zeros in a convex domain C\mathfrak {C} that contains the origin. Let the polynomial g(z)g(z) of the same degree have only real zeros which lie in the interval −1,0-1, 0. The polynomial h(z)h(z) of degree nn derived from the two polynomials by composition [151] then also has all its zeros in C\mathfrak {C}.

?
Problem 5.154

If all the zeros of the polynomial

f(z)=a0+(n1)a1z+(n2)a2z2+⋯+(nn−1)an−1zn−1+anz f(z) = a_0+\binom {n}{1}a_1z+\binom {n}{2}a_2z^2+\cdots +\binom {n}{n-1}a_{n-1}z^{n-1}+a_nz

of degree nn lie in the real interval −a,a-a, a, and those of the polynomial

g(z)=b0+(n1)b1z+(n2)b2z2+⋯+(nn−1)bn−1zn−1+bnzn g(z) = b_0+\binom {n}{1}b_1z+\binom {n}{2}b_2z^2+\cdots +\binom {n}{n-1}b_{n-1}z^{n-1}+b_nz^n

of degree nn in the interval −b,0-b, 0 (or 0,b0, b) then the zeros of the polynomial

h(z)=a0b0+(n1)a1b1z+(n2)a2b2z2+⋯+(nn−1)an−1bn−1zn−1+anbnzn h(z) = a_0b_0+\binom {n}{1}a_1b_1z+\binom {n}{2}a_2b_2z^2+\cdots +\binom {n}{n-1}a_{n-1}b_{n-1}z^{n-1}+a_nb_nz^n

of degree nn lie in the interval −ab,ab-ab, ab. (a,ba, b are positive numbers.)

?
Problem 5.155

Let all the zeros of the polynomial

a0+a1z+a2z2+⋯+anzn a_0+a_1z+a_2z^2+\cdots +a_nz^n

of degree nn be real and those of the polynomial

b0+b1z+b2z2+⋯+bnzn b_0+b_1z+b_2z^2+\cdots +b_nz^n

of degree nn be real and of the same sign. Then all the zeros of the polynomial

a0b0+a1b1z+a2b2z2+⋯+anbnzn a_0b_0+a_1b_1z+a_2b_2z^2+\cdots +a_nb_nz^n

are also real. (z=∞z=\infty counts as a real zero.)

?
Problem 5.156

Under the assumptions of 155, all the zeros of the polynomial

a0b0+1!a1b1z+2!a2b2z2+⋯+n!anbnzn a_0b_0+1!a_1b_1z+2!a_2b_2z^2+\cdots +n!a_nb_nz^n

are real.

?