J. Math. Anal. Appl.

336 (2007) 14561465

www.elsevier.com/locate/jmaa
An L
p
inequality for polynomials
M.A. Qazi
Department of Mathematics, Tuskegee University, Tuskegee, AL 36088, USA
Received 26 November 2006
Available online 28 March 2007
Submitted by Richard M. Aron
Abstract
Let P
n
be the class of all polynomials of degree at most n, and let M
p
(g; ) denote the L
p
mean of g on
the circle of radius centered at the origin. We specify a number

(0, 1), depending on n and k, such

that for any f P
n
, the ratio M
p
(f
(k)
; )/M
p
(f ; 1) is maximized by f (z) := z
n
for all [

, )
and p 1. The interest of the result lies in the fact that

is strictly less than 1.

2007 Elsevier Inc. All rights reserved.
Keywords: Polynomials; Bernsteins inequality; Zygmunds inequality
1. Introduction and statement of results
For any entire function , let
M
p
(; r) :=
_
1
2

_
re
i
_

p
d
_
1/p
(0 <p <, r >0). (1)
By a classical result of G.H. Hardy [4], M
p
(; r) is a strictly increasing function of r except in
the case where is a constant. It is well known (see, for example, [5, p. 143]) that for any given
r >0,
M
p
(; r) max
|z|=r

(z)

as p .
E-mail address: qazima@aol.com.
0022-247X/$ see front matter 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jmaa.2007.03.065
M.A. Qazi / J. Math. Anal. Appl. 336 (2007) 14561465 1457
Hence,
M(; r) := max
|z|=r

(z)

(r >0)
may be seen as M

(; r). We add that, by the maximum modulus principle, M

(; r) is a
strictly increasing function of r unless f is a constant.
Let P
n
be the class of all polynomials of degree at most n. It is clear that if f belongs to P
n
then so does f

(z) :=z
n
f (1/z), and so for R >1 and any p >0, we have
M
p
(f ; R) =R
n
M
p
_
f

;
1
R
_
R
n
M
p
_
f

; 1
_
=R
n
M
p
(f ; 1),
where equality holds in the inequality only if f

is a constant. Hence,
M
p
(f ; R) <R
n
M
p
(f ; 1) (0 <p ; 1 <R <), (2)
unless f is a constant multiple of z
n
.
It is also known that for any polynomial f of degree at most n, we have
M
p
(f

; 1) nM
p
(f ; 1) (0 <p ). (3)
The inequality is sharp. It becomes an equality for f (z) :=cz
n
.
The case p = of (3) is known as S. Bernsteins inequality (see [8, Chapter 14]). It was
proved by A. Zygmund [10] for p [1, ). The fact that (3) is also true for 0 <p <1 is a result
of V.V. Arestov [1].
The following result is an immediate consequence of (2) and (3).
Theorem A. Let f be a polynomial of degree at most n. Then
M
p
(f

; ) n
n1
M
p
(f ; 1) (0 <p ; 1 <). (4)
The inequality is sharp. It becomes an equality for f (z) :=cz
n
.
We show that the restriction on in (4) can be relaxed. In order to state our result we need to
introduce some notation.
Let
A
m
() :=

1
m1
m

1
m
m1
0
m1
m
1
2
m
m2
1
m
m1
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
m
m1
2
m
m2
1
m1
m
0
1
m
m1

m1
m
1

(m+1)(m+1)
, (5)
and denote by det A
m
() the determinant of this matrix. It may be noted that det A
m
() is a poly-
nomial in 1/ and so a continuous function of for 0 < <.
Notation 1. Let
m
=
m,m+1
be the largest root of det A
m
() = 0 in (0, 1).
It is clear that det A
m
() is a polynomial of degree m
2
1 in 1/. For large m, it may be hard
to compute
m
explicitly. We would have liked to give an asymptotic estimate for it as m tends
to innity, but at this time we do not have one.
We shall see that det A
m
() is strictly positive in (
m
, ). Now, we are ready to state our
extension of Theorem A.
1458 M.A. Qazi / J. Math. Anal. Appl. 336 (2007) 14561465
Theorem 1. Let M

(; r) := max
|z|=r
|(z)|, and for 0 < p < , let M
p
(; r) be as in (1).
Furthermore, as indicated in Notation 1, let
m
be the largest root of det A
m
() = 0 in (0, 1).
Then, for any polynomial f of degree at most n, and 1 p , we have
M
p
_
f
(k)
;
_

n!
(n k)!

nk
M
p
(f ; 1) (k = 1, . . . , n) (6)
if

k1
j=0

nj
<.
Remark. Let f (z) :=z
n
+z
n1
. Then M

(f ; 1) = 2 and
M

_
f
(k)
;
_
=
n!
(n k)!

nk
+
(n 1)!
(n k 1)!

nk1
(k = 1, . . . , n 1).
Hence,
M

_
f
(k)
;
_
>
n!
(n k)!

nk
M

(f ; 1)
for 0 <1 k/n. This shows that

k1
j=0

nj
cannot be smaller than 1 k/n.
In the case where p = 2 we can easily prove the following result which leaves little to be
desired.
Theorem 2. For any polynomial f (z) :=

n
=0
a

of degree at most n and k = 1, . . . , n, we

have
_
M
2
_
f
(k)
;
__
2
+
_
n!
(n k)!
_
2

2n2k
k1

=0
|a

|
2

_
n!
(n k)!
_
2

2n2k
_
M
2
(f ; 1)
_
2
(7)
if 1 k/n. The inequality may not hold for any < 1 k/n. Besides, the coefcient of

k1
=0
|a

|
2
cannot be replaced by a larger number.
Inequality (7) says in particular that
M
2
(f

; ) n
n1
_
_
M
2
(f ; 1)
_
2
|a
0
|
2
_
1
1
n
_
.
2. Some auxiliary results
We start with a standard notation.
For a function g that is dened and bounded on the unit circle we shall sometimes write g

to denote sup{|g(z)|: |z| = 1}.

Next, we recall the notion of Hadamard product of two holomorphic functions dened in
concentric disks.
Let A(z) =

=0

and B(z) =

=0

be holomorphic in the disks |z| < R

1
and
|z| < R
2
, respectively. Then,

=0

denes a function holomorphic in the disk |z| <

R
1
R
2
. It is called the Hadamard product of A and B, for which the usual notation is A B.
With this we are ready to introduce the following denition. It is taken from [8, p. 409].
Denition 1. A polynomial Q of degree at most n is said to belong to the class B
0
n
if Q(0) = 1
and Q f

1 for every polynomial f of degree at most n such that f

1.
M.A. Qazi / J. Math. Anal. Appl. 336 (2007) 14561465 1459
Whether a polynomial Q belongs to B
0
n
or not can be decided by studying the deniteness of
an associated Hermitian form.
Denition 2. A Hermitian form
H(x, x) :=
n

,=1
h

is called positive semidenite if H(x, x) 0 for arbitrary values of the variables x

1
, . . . , x
n
, not
all equal to zero; it is called positive denite if H(x, x) > 0 for arbitrary values of the variables
x
1
, . . . , x
n
, not all equal to zero.
Here is then [8, Theorem 12.2.8] a necessary and sufcient condition for a polynomial Q to
belong to B
0
n
.
Lemma 1. The polynomial Q(z) := 1 +

n
=1
c

belongs to the class B

0
n
introduced in Deni-
tion 1, if and only if the Hermitian form
x :=(x
0
, . . . , x
n
)

,
c

(c
0
= 1, c
k
=c
k
, 1 k n)
is positive semidenite.
Let us recall that a principal minor of a matrix A is a minor whose diagonal is part of the
diagonal of A.
The deniteness of a Hermitian form can be decided in terms of the principal minors of
its matrix. The precise relationship is summarized in Lemma 2 ([3, p. 337], see Theorems 19
and 20). An alternative reference could be [6, Chapter XIII]. However, as a cautionary note, it
may be mentioned that in [6] the denition of semidenite is slightly different.
Lemma 2.
(i) A Hermitian form
H(x, x) :=
n

,=1
h

is positive denite if and only if its leading principal minors are all positive, that is
D
k
:=

h
11
h
1k



h
k1
h
kk

>0 (k = 1, . . . , n).
(ii) A Hermitian form is positive semidenite if and only if all its principal minors are non-
negative.
The next lemma, which is a result of E. Fischer [2], simplies certain things for us in a
signicant way. For its proof we refer the reader to [6, p. 420].
1460 M.A. Qazi / J. Math. Anal. Appl. 336 (2007) 14561465
Lemma 3. Let (c
jk
) be a positive denite Hermitian (n + 1) (n + 1) matrix. Then, for =
1, . . . , n, we have

c
1,1
c
1,



c
,1
c
,

c
+1,+1
c
+1,n+1



c
n+1,+1
c
n+1,n+1

c
1,1
c
1,n+1



c
n+1,1
c
n+1,n+1

.
For the proof of Theorem 1, we shall also need the following auxiliary result [7, Lemma 3].
Lemma 4. Let P
n
be the linear space of all polynomials of degree at most n, normed by f

:=
max
|z|=1
|f (z)|. Also let t
0
, . . . , t
n
be an arbitrary set of n + 1 numbers in C, and denote by L
the linear functional dened on P
n
, by
f t
0
a
0
+ +t
n
a
n
_
f (z) :=
n

=0
a

_
.
Furthermore, let N := L. Then, for any non-decreasing convex function on [0, ), we have

_
|

n
=0
t

e
i
|
N
_
d

=0
a

e
i

_
d.
3. Proofs
Proof of Theorem 1. By Theorem A, inequality (6) holds for all p (0, ] and any 1. So,
we only need to prove that it also holds for

k1
j=0

nj
<1 if p [1, ].
First we shall treat the case where p = and k = 1. Thus, with
m
as in Notation 1, we wish
to prove that if
q(z) =q
n,
(z) := 0 +z +2z
2
+ +
1
z

+ +n
n1
z
n
( >0),
then, for any polynomial f of degree at most n, we have
q f

:= max
|z|=1

(q f )(z)

n
n1
f

(
n
<1). (8)
If f belongs to P
n
then so does f

(z) :=z
n
f (1/z). Besides, f

= f

. Hence, (8) holds

if and only if
_
_
q f

_
_

n
n1
f

(
n
<1; f P
n
). (9)
Since q f

= (q f

and (q f

=q

f inequality (9) is equivalent to

_
_
_
q

f
__
_

n
n1
f

(
n
<1; f P
n
).
Thus, setting
Q

(z) :=
1
n
n1
q

(z) = 1 +
n

=1
(n )
n1
n
n1
z

we have to prove that

Q

1 (
n
<1; f P
n
, f

1).
M.A. Qazi / J. Math. Anal. Appl. 336 (2007) 14561465 1461
Since Q

(0) = 1, this means that Q belongs to the class B

0
n
introduced in Denition 1. Hence, if
c
0
=c
0
() := 1, and c

() =c

() :=
(n )
n1
n
n1
for = 1, . . . , n,
then, in view of Lemma 1, it would sufce to show that the Hermitian form
(x
0
, . . . , x
n
) =

,
c

is positive denite for

n
< 1. This is exactly what we shall do.
We shall show that the leading principal minors of det A
n
() are all strictly positive for
n
<
1. Let us start by evaluating det A
n
(1). Note that
n
n+1
det A
n
(1) :=

n (n 1) 1 0
(n 1) n 2 1
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1 2 n (n 1)
0 1 (n 1) n

(n+1)(n+1)
.
In order to evaluate this latter determinant we may perform the following operations one after
the other.
(a) Subtract the (j 1)th row from the jth row for j =n +1, n, . . . , 2;
(b) add the 2nd row to the last row;
(c) add the last column to the rst;
(d) expand by the last row.
We obtain n
n+1
det A
n
(1) = 2
1
, where

1
=

n (n 1) (n 2) 2 1
0 1 1 1 1
0 1 1 1 1
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
0 1 1 1 1
0 1 1 1 1

nn
.
Expanding
1
by its rst column we write
1
= n
2
. In order to evaluate
2
, we may add the
last column to the rst, and then we readily see that
2
= 2
n2
. Thus, n
n+1
det A
n
(1) = n2
n1
,
that is
det A
n
(1) =
2
n1
n
n
.
Recall that A
n
() is an (n + 1) (n + 1) matrix. For = 1, . . . , n + 1, its leading principal
minor of order is the determinant of
A
n,
() :=

1
n1
n

n+2
n
2
n+1
n
1
n1
n
1
n+3
n
3
n+2
n
2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
n+2
n
2
1
n1
n
n+1
n
1
n+2
n
2

n1
n
1

.
1462 M.A. Qazi / J. Math. Anal. Appl. 336 (2007) 14561465
Note that A
n,n+1
() is simply A
n
().
As we have already seen,
det A
n,n+1
(1) = det A
n
(1) =
2
n1
n
n
.
The minors det A
n,1
(1), . . . , det A
n,n
(1) can be evaluated the same way as det A
n,n+1
(1). So,
without providing any additional details we leave it to the reader to check that
det A
n,
(1) =
2n +1
n

2
2
( = 1, . . . , n +1).
Now, note that det A
n,1
(), . . . , det A
n,n+1
() are continuous functions of in (0, ), and we
know that they are all strictly positive for = 1. Hence, for each {1, . . . , n +1}, there exists
a positive number
n,
(0, 1) such that det A
n,
() >0 for
n,
< 1 and det A
n,
(
n,
) = 0.
We claim that

n,n+1

n,
( = 1, . . . , n). (10)
For this we rst note that if A
n
() is written as an (n +1) (n +1) matrix (c
jk
), then

c
+1,+1
c
+1,n+1



c
n+1,+1
c
n+1,n+1

=A
n,n+1
() ( = 1, . . . , n).
For a proof (of (10)) by contradiction let us suppose that (10) is not true, i.e. there exists an
integer {1, . . . , n} such that

n,
= max
1n

n,
>
n,n+1
, (11)
and so, by part (i) of Lemma 2, the matrix A
n
() is positive denite for
n,
< 1. Then,
Lemma 3 in conjunction with (11) implies that
det A
n,
() det A
n,n+1
() det A
n,n+1
() (
n,
< 1). (12)
As
n,
the left-hand side of (12) tends to 0 whereas its right-hand side tends to
det A
n,n+1
(
n,
)a strictly positive quantity if (11) was really true. This is a contradiction
which proves (10). Hence, the matrix A
n
() is positive denite for
n
< 1 and, by part (ii)
of Lemma 2, all the principal minors of A
n
() are non-negative for these values of . By con-
tinuity, they remain non-negative for =
n
. Once again, part (ii) of Lemma 2 implies that the
matrix A
n
() is positive semidenite for
n
1. Applying Lemma 1 we conclude that
M

(f

; ) n
n1
M

(f ; 1) (13)
for
n
1, which completes the proof of (6) in the case where p = and k = 1.
Now, let us turn to the case where p [1, ). For any given [
n
, 1] let
t

:=
1
( = 0, 1, . . . , n).
Then
f t
0
a
0
+ +t
n
a
n
_
f (z) :=
n

=0
a

_
M.A. Qazi / J. Math. Anal. Appl. 336 (2007) 14561465 1463
denes a linear functional L on the linear space P
n
of all polynomials f of degree at most n.
By (13), the norm L of this functional is n
n1
. Lemma 4 with N = n
n1
shows that (6)
holds also for p [1, ) if k = 1.
Now we shall show how to prove (6) for k = 2, etc.
Let 1 p <. We already know that
_
1
2

n
e
i
_

p
d
_
1/p
n
n1
n
_
1
2

f
_
e
i
_

p
d
_
1/p
.
So, applying the case k = 1 of (6) to the polynomial f

(
n
z), which is a polynomial of degree at
most n 1, we see that for
n1
< we have

n
_
1
2

n
e
i
_

p
d
_
1/p
(n 1)
n2
_
1
2

n
e
i
_

p
d
_
1/p
n(n 1)
n1
n

n2
_
1
2

f
_
e
i
_

p
d
_
1/p
.
Thus, for any p [1, ) the inequality
_
1
2

_
e
i
_

p
d
_
1/p
n(n 1)
n2
_
1
2

f
_
e
i
_

p
d
_
1/p
(14)
holds if
n

n1
< . This proves (6) in the case where k = 2 and 1 p < . Letting
p in (14), we see that (6) holds also when p = . Repeated application of this argument
gives (6) for any k and any p [1, ]. 2
Proof of Theorem 2. Since
f
(k)
(z) =
nk

=0
(n )!
(n k )!
a
n
z
nk
,
we have
_
M
2
_
f
(k)
;
__
2
=
nk

=0
_
(n )!
(n k )!
_
2
|a
n
|
2

2(nk)
=
_
n!
(n k)!
_
2
nk

=0
_
(n )!
(n k )!
(n k)!
n!
_
2
|a
n
|
2

2(nk)
.
Now, note that
(n )!
(n k )!
(n k)!
n!
=

1
=0
(n k )

1
=0
(n )
=
1

=0
_
1
k
n
_

_
1
k
n
_

.
Hence,
_
(n )!
(n k )!
(n k)!
n!
_
2

2(nk)

2(nk)
_
n k
n
<
_
,
1464 M.A. Qazi / J. Math. Anal. Appl. 336 (2007) 14561465
which implies that if 1
k
n
, then
_
M
2
_
f
(k)
;
__
2

_
n!
(n k)!
_
2

2(nk)
nk

=0
|a
n
|
2
.
From this (7) readily follows.
We claim that (7) may not hold for 0 < <1
k
n
. In fact,
_
(n 1)!
(n k 1)!
_
2

2n2k2
>
_
n!
(n k)!
_
2

2n2k
_
0 < <1
k
n
_
,
and so for any polynomial of the form f (z) := a
n
z
n
+ a
n1
z
n1
+

k1
=0
a

with a
n1
= 0
and k n 1, we have
_
M
2
_
f
(k)
;
__
2
+
_
n!
(n k)!
_
2

2n2k
k1

=0
|a

|
2
=
_
n!
(n k)!
_
2
|a
n
|
2

2n2k
+
_
(n 1)!
(n k 1)!
_
2
|a
n1
|
2

2n2k2
+
_
n!
(n k)!
_
2

2n2k
k1

=0
|a

|
2
>
_
n!
(n k)!
_
2

2n2k
_
M
2
(f ; 1)
_
2
_
0 < <1
k
n
_
.
In order to see that the coefcient {(n!)/(n k)!}
2

2n2k
of

k1
=0
|a

|
2
in (7) cannot be
replaced by a larger quantity, we may consider any polynomial of the form f (z) := a
n
z
n
+

k1
=0
a

with

k1
=0
|a

| >0. 2
4. The restriction on p in Theorem 1
It is known (see [5, p. 139]) that if M
p
(; r) is as in (1), then for any given r >0,
M
p
(; r) exp
_
1
2

log

_
re
i
_

d
_
as r 0.
This is the motivation behind the denition
M
0
(; r) := exp
_
1
2

log

_
re
i
_

d
_
(0 <r <).
Using Jensens theorem [9, p. 124] it can be shown that if f is a polynomial of the form
f (z) :=a
m

m
=1
(z z

), then
exp
_
1
2

log

f
_
e
i
_

d
_
= |a
m
|
m

=1
max
_
|z

|, 1
_
,
M.A. Qazi / J. Math. Anal. Appl. 336 (2007) 14561465 1465
and so
M
0
(f ; 1) = |a
m
|
m

=1
max
_
|z

|, 1
_
. (15)
Now, let f (z) := (z 1)
n
. Then, we may apply (15) to conclude that M
0
(f ; 1) = 1. The
same formula shows that
M
0
(f

; ) =n
n1
n1

=1
max
_

1
, 1
_
=n (0 < <1),
that is M
0
(f

; ) = nM
0
(f ; 1) for 0 < < 1. So, there is no chance of (6) being true for all
p [0, 1).
Note added in proof
Within the last four weeks I have learned that the case p = of Theorem 1 appeared rst
in a paper entitled On a class of extremal problems for polynomials in the unit circle by H.S.
Shapiro, published in Portugaliae Mathematica 20 (1961) 6793 (see Theorem 9 on page 91).
However, our proof is totally diferent; it has nothing in common with his.
References
[1] V.V. Arestov, On integral inequalities for trigonometric polynomials and their derivatives, Math. USSR Izv. 18
(1982) 117; Izv. Akad. Nauk SSSR, Ser. Mat. 45 (1981) 322.
[2] E. Fischer, ber den Hadamardschen Determinantensatz, Arch. Math. Phys. (3) 13 (1908) 3240.
[3] F.R. Gantmacher, The Theory of Matrices, vol. I, Chelsea, New York, 1959.
[4] G.H. Hardy, On the mean value of the modulus of an analytic function, Proc. London Math. Soc. (2) 14 (1915)
269277.
[5] G.H. Hardy, J.E. Littlewood, G. Plya, Inequalities, Cambridge Univ. Press, 1934.
[6] L. Mirsky, An Introduction to Linear Algebra, Dover, New York, 1990.
[7] Q.I. Rahman, Functions of exponential type, Trans. Amer. Math. Soc. 135 (1969) 295309.
[8] Q.I. Rahman, G. Schmeisser, Analytic Theory of Polynomials, London Math. Soc. Monogr. (N.S.), vol. 26, Claren-
don Press, Oxford, 2002.
[9] E.C. Titchmarsh, The Theory of Functions, second ed., Oxford Univ. Press, 1939.
[10] A. Zygmund, A remark on conjugate series, Proc. London Math. Soc. (2) 34 (1932) 392400.