TURÁN-TYPE INEQUALITIES FOR SOME q-SPECIAL FUNCTIONS

(1)

Turán-Type Inequalities Kamel Brahim vol. 10, iss. 2, art. 50, 2009

Title Page

Contents

JJ II

J I

Page1of 11 Go Back Full Screen

Close

TURÁN-TYPE INEQUALITIES FOR SOME q-SPECIAL FUNCTIONS

KAMEL BRAHIM

Institut Préparatoire aux Études d’Ingénieur de Tunis Tunis

EMail:kamel710@yahoo.fr

Received: 24 March, 2008

Accepted: 10 April, 2009

Communicated by: S.S. Dragomir 2000 AMS Sub. Class.: 26D07, 26D15, 33D05.

Key words: Turan’s inequality,q-polygamma functions,q-zeta function.

Abstract: In this paper, we give new Turán-type inequalities for someq-special functions, using aq- analogue of a generalization of the Schwarz inequality.

(2)

Title Page Contents

JJ II

J I

Close

1. Introduction

In [9], P. Turán proved that the Legendre polynomialsP_n(x)satisfy the inequality (1.1) P_n+1² (x)−P_n(x)P_n+2(x)≥0, x∈[−1,1], n= 0,1,2, . . .

and equality occurs only ifx=±1.

This inequality been the subject of much attention and several authors have provided new proofs, generalizations, extensions and refinements of (1.1).

In [7], A. Laforgia and P. Natalini established some new Turán-type inequalities for polygamma and Riemann zeta functions:

Theorem 1.1. For n = 1,2, . . . we denote by ψn(x) = ψ⁽ⁿ⁾(x) the polygamma functions defined as then−thderivative of the psi function

ψ(x) = Γ⁰(x)

Γ(x), x >0

with the usual notation for the gamma function. Then ψm(x)ψn(x)≥ψ²m+n

2

(x),

where ^m+n₂ is an integer

Theorem 1.2. We denote byζ(s)the Riemann zeta function. Then

(s+ 1) ζ(s)

ζ(s+ 1) ≥sζ(s+ 1)

ζ(s+ 2), ∀s >1.

(1.2)

The main aim of this paper is to give some new Turán-type inequalities for the q-polygamma and q-zeta [2] functions by using a q-analogue of the generalization of the Schwarz inequality.

(4)

Title Page Contents

JJ II

J I

Close

To make the paper more self contained we begin by giving some usual notions and notations used inq-theory. Throughout this paper we will fixq∈]0,1[and adapt the notations of the Gasper-Rahman book [4].

Letabe a complex number, theq-shifted factorial are defined by:

(1.3) (a;q)₀ = 1; (a;q)_n =

n−1

Y

k=0

(1−aq^k) n= 1,2, . . .

(1.4) (a;q)∞= lim

n→+∞(a;q)_n=

∞

Y

k=0

(1−aq^k).

Forxcomplex we denote

(1.5) [x]_q = 1−q^x

1−q .

Theq-Jackson integrals from0toaand from0to∞are defined by [4,5]:

(1.6)

Z a

0

f(x)d_qx= (1−q)a

∞

X

n=0

f(aqⁿ)qⁿ

and (1.7)

Z ∞

0

f(x)d_qx= (1−q)

∞

X

n=−∞

f(qⁿ)qⁿ provided the sums converge absolutely.

Jackson [5] defined theq-analogue of the Gamma function as:

(1.8) Γ_q(x) = (q;q)∞

(q^x;q)∞

(1−q)^1−x x6= 0,−1,−2, . . . .

(5)

Title Page Contents

JJ II

J I

Close

It satisfies the functional equation:

(1.9) Γq(x+ 1) = [x]qΓq(x), Γq(1) = 1 and tends toΓ(x)whenqtends to 1.

Moreover, it has theq-integral representation (see [1,3]) Γ_q(s) = K_q(s)

Z ∞

0

x^s−1e^−x_q d_qx, where

e^x_q = 1 ((1−q)x;q)∞

, and

K_q(t) = (1−q)^−s

1 + (1−q)⁻¹ · (−(1−q), q)∞(−(1−q)⁻¹, q)∞

(−(1−q)q^s, q)_∞(−(1−q)⁻¹q^1−s, q)_∞.

Lemma 1.3. Leta∈R+∪ {∞}and letf andgbe two nonnegative functions. Then

(1.10) Z a

0

g(x)f^m+n² (x)dqx 2

≤ Z a

0

g(x)f^m(x)dqx

Z a

0

g(x)fⁿ(x)dqx

,

wheremandnbelong to a setSof real numbers, such that the integrals (1.10) exist.

Proof. Lettinga >0, by definition of theq-Jackson integral, we have Z a

0

g(x)f^m+n² (x)d_qx= (1−q)a

∞

X

p=0

g(aq^p)f^m+n² (aq^p)q^p (1.11)

= lim

N→+∞(1−q)a

N

X

p=0

g(aq^p)f^m+n² (aq^p)q^p

(6)

Title Page Contents

JJ II

J I

Close

By the use of the Schwarz inequality for finite sums, we obtain

(1.12)

N

X

p=0

g(aq^p)f^m+n² (aq^p)q^p

!²

≤

N

X

p=0

g(aq^p)f^m(aq^p)q^p

! _N X

p=0

g(aq^p)fⁿ(aq^p)q^p

! .

The result follows from the relation (1.11) and (1.12).

To obtain the inequality for a = ∞, it suffices to write the inequality (1.10) for a=q^−N, then tendN to∞.

(7)

Title Page Contents

JJ II

J I

Close

2. The q-Polygamma Functions

Theq-analogue of the psi functionψ(x) = ^Γ_Γ(x)⁰^(x) is defined as the logarithmic derivative of theq-gamma function,ψ_q(x) = ^Γ

0 q(x) Γq(x). From (1.8), we get forx >0

ψ_q(x) = −Log(1−q) + Logq

∞

X

n=0

q^n+x 1−q^n+x

=−Log(1−q) + Logq

∞

X

n=1

q^nx 1−qⁿ.

The last equality implies that

(2.1) ψ_q(x) = −Log(1−q) + Logq 1−q

Z q

0

t^x−1 1−td_qt.

Theorem 2.1. Forn = 1,2, . . ., putψ_q,n = ψq⁽ⁿ⁾ then-th derivative of the function ψq. Then

(2.2) ψ_q,n(x)ψ_q,m(x)≥ψ_q,²m+n 2

(x),

where ^m+n₂ is an integer.

Proof. Letmandnbe two integers of the same parity.

From the relation (2.1) we deduce that ψ_q,n(x) = Logq

1−q Z q

0

(Logt)ⁿt^x−1 1−t d_qt.

(8)

Title Page Contents

JJ II

J I

Close

Applying Lemma1.3withg(t) = ^t_1−t^x−1, f(t) = (−Logt)anda=q, we obtain (2.3)

Z q

0

t^x−1

1−t(−Logt)ⁿd_qt Z q

0

t^x−1

1−t(−Logt)^md_qt

≥ Z q

0

t^x−1

1−t(−Logt)^m+n² d_qt 2

,

which gives, sincem+nis even,

(2.4) ψ_q,n(x)ψ_q,m(x)≥ψ_q,²m+n 2

(x).

Takingm=n+ 2, one obtains:

Corollary 2.2. For allx >0we have

(2.5) ψ_q,n(x)

ψ_q,n+1(x) ≥ ψ_q,n+1(x)

ψ_q,n+2(x), n = 1,2, . . . .

(9)

Title Page Contents

JJ II

J I

Close

3. The q- Zeta function

Forx >0, we put

α(x) = Log(x) Log(q) −E

Log(x) Log(q)

and

{x}_q = [x]_q q^x+α([x]^q⁾, whereE

Log(x) Log(q)

is the integer part of ^Log(x)_Log(q).

In [2], the authors defined theq-Zeta function as follows

(3.1) ζ_q(s) =

∞

X

n=1

1 {n}^s_q =

∞

X

n=1

q^(n+α([n]^q^))s [n]^s_q .

They proved that it is aq-analogue of the classical Riemann Zeta function and we have for alls ∈Csuch that<(s)>1,

ζq(s) = 1 Γe_q(s)

Z ∞

0

t^s−1Zq(t)dqt,

where for allt >0,

Z_q(t) =

∞

X

n=1

e^−{n}_q ^q^t and eΓ_q(t) = Γ_q(t) K_q(t). Theorem 3.1. For alls >1we have

(3.2) [s+ 1]_q ζ_q(s)

ζq(s+ 1) ≥q[s]_qζ_q(s+ 1) ζq(s+ 2).

(10)

Title Page Contents

JJ II

J I

Close

Proof. Fors >1the functionq-zeta satisfies the following relation

(3.3) ζq(s) = 1

Γe_q(s) Z ∞

0

t^s−1Zq(t)dqt.

Applying Lemma1.3withg(t) =Z_q(t),f(t) =twe obtain (3.4)

Z ∞

0

t^s−1Z_q(t)d_qt Z ∞

0

t^s+1Z_q(t)d_qt≥ Z ∞

0

t^sZ_q(t)d_qt 2

.

Further, using (3.3), this inequality becomes

(3.5) ζ_q(s)eΓ_q(s)ζ_q(s+ 2)eΓ_q(s+ 2)≥[ζ_q(s+ 1)]²h

eΓ_q(s+ 1)i2

.

So, by using the relationΓeq(s+ 1) =q^−s[s]qeΓq(s),we obtain (3.6) [s+ 1]_qζ_q(s)ζ_q(s+ 2) ≥q[s]_q[ζ_q(s+ 1)]² which completes the proof.

(11)

Title Page Contents

JJ II

J I

Close

References

[1] A.De SOLE AND V.G. KAC, On integral representations of q-gamma and q- beta function, Atti Accad. Naz. Lincei Cl. Sci. Fis. Mat. Natur. Rend. Lincei Mat.

Appl., 16(9) (2005), 11–29.

[2] A. FITOUHI, N. BETTAIBIANDK. BRAHIM, The Mellin transform in quan- tum calculus, Constructive Approximation, 23(3) (2006), 305–323.

[3] A. FITOUHI AND K. BRAHIM, Tauberian theorems in quantum calculus, J.

Nonlinear Mathematical Physics, 14(3) (2007), 316–332.

[4] G. GASPERANDM. RAHMAN, Basic hypergeometric series, Encyclopedia of Mathematics and its Applications, Vol. 35, Cambridge Univ. Press, Cambridge, UK, 1990.

[5] F.H. JACKSON, On aq-definite integral, Quarterly J. Pure and Appl. Math., 41 (1910), 193–203.

[6] V.G. KACANDP. CHEUNG, Quantum Calculus, Universitext, Springer-Verlag, New York, 2002.

[7] A. LAFORGIA AND P. NATALINI, Turán -type inequalities for some special functions, J. Ineq. Pure. App. Math., 7(1) (2006), Art. 22. [ONLINE: http:

//jipam.vu.edu.au/article.php?sid=638]

[8] G. SZEG ˝O, On an inequality of P. Turán concerning Legendre polynomials, Bull.

Amer. Math. Soc., 54 (1948), 401–405.

[9] P. TURÁN, On the zeros of the polynomials of Legendre, Casopis pro Pestovani Mat. a Fys, 75 (1950), 113–122.