Int. Journal of Math. Analysis, Vol. 2, 2008, no.

13, 629 - 638

The Hadamards Inequality for sConvex

Function of 2-Variables on the Co-ordinates
M. Alomari and M. Darus

School of Mathematical Sciences

Faculty of Science and Technology
Universiti Kebangsaan Malaysia
Bangi 43600 Selangor, Malaysia
alomari@math.com

Corresponding author. maslina@pkrisc.cc.ukm.my

Abstract
In this paper the extension of Hadamards type inequality for s
convex function and sconvex functions on the co-ordinates defined in
2-variables and some applications are given.

Keywords: sHadamards inequality, sConvex function, Jensens in-

equality

1 Introduction
Let f : I R R be a convex mapping dened on the interval I of real
numbers and a, b I, with a < b. The following double inequality:
  b
a+b 1 f (a) + f (b)
f f (x) dx (1)
2 ba a
2

is known in the literature as Hadamards inequality for convex mappings.

In [1] Hudzik and Maligrada considered among others the class of functions
which are sconvex in the second sense. This class is dened in the following
way: a function f : [0, ) R is said to be sconvex in the second sense if
f (x + (1 ) y) s f (x) + (1 )s f (y) (2)
holds for all x, y [0, ), [0, 1] and for some xed s (0, 1]. It can be
easily seen that every 1convex function is convex.
630 M. Alomari and M. Darus

In [2] Dragomir and Fitzpatrick proved a variant of Hadamards inequality

which holds for sconvex functions in the second sense.

Theorem A

Suppose that f : [0, ) [0, ) is an sconvex function in the second

sense, where s (0, 1) and let a, b [0, ), a < b. If f L1 [0, 1], then the
following inequalities hold:
  b
a+b 1 f (a) + f (b)
2 s1
f f (x) dx (3)
2 ba a
s+1
1
the constant k = s+1 is the best possible in the second inequality in (1.3). The
above inequalities are sharp.

In [3], Dragomir established the following similar inequality of Hadamard-

type for co-ordinated convex mapping on a rectangle from the plane R2 .

Precisely, if f : [a, b] [c, d] R is convex function one can dene the

following mapping on [0, 1]2 such as:
b d  
1 a+b c+d
H (t, r) = f tx + (1 t) , ry + (1 r) dydx
(b a) (d c) a c
2 2

Theorem B

Suppose that f : R is co-ordinated convex on . Then one has the

inequalities
  b d
a+b c+d 1
f , f (x, y) dydx
2 2 (b a) (d c) a c
f (a, c) + f (a, d) + f (b, c) + f (b, d)
(4)
4
The above inequalities are sharp.

Also, one can has the following properties for H (see [3-5]):

(i) H is coordinated convex on [0, 1]2.

(ii) One has the bounds
b d
1
sup H (t, r) = f (x, y) dxdy = H (1, 1)
(t,r)[0,1]2 (b a) (d c) a c
Hadamards inequality 631

and
 
a+b c+d
inf H (t, r) = f , = H (0, 0) .
(t,r)[0,1]2 2 2

2 Hadamards Inequality
First of all, let us start with the following denition:

Definition 2.1

Consider the bidimensional interval := [a, b] [c, d] in [0, )2 with a < b

and c < d. The mapping f : R is sconvex on if

f (x + (1 ) z, y + (1 ) w) s f (x, y) + (1 )s f (z, w)

holds for all (x, y), (z, w) with [0, 1] and for some xed s (0, 1].

A function f : R is sconvex on is called co-ordinated sconvex on

if the partial mappings fy : [a, b] R, fy (u) = f (u, y) and fx : [c, d] R,
fx (v) = f (x, v), are sconvex for all y [c, d] and x [a, b] with some xed
s (0, 1].

Lemma 2.1

Every sconvex mapping f : = [a, b] [c, d] [0, )2 [0, ) is s

convex on the co-ordinates, but the converse is not true in general.

Proof.

Suppose that f : = [a, b] [c, d] [0, )2 [0, ) is sconvex on .

Consider the function fx : [c, d] [0, ), fx (v) = f (x, v). Then for [0, 1]
and v1 , v2 [c, d], one has:

fx (v1 + (1 ) v2 ) = f (x, v1 + (1 ) v2 )
= f (x + (1 ) x, v1 + (1 ) v2 )
s f (x, v1 ) + (1 )s f (x, v2 )
= s fx (x, v1 ) + (1 )s fx (x, v2 ) .

Therefore, fx (v) = f (x, v) is sconvex on [c, d] .

632 M. Alomari and M. Darus

The fact that fx : [a, b] [0, ), fy (u) = f (u, y) is also sconvex on [a, b]
for all y [c, d] goes likewise and we shall omit the details.

In [3] Dragomir gave a mapping f0 : [0, 1]2 [0, ) dened by f0 (x, y) =

xy which is convex on the coordinates but is not convex. We consider the
same function with s = 1 to prove that the sconvexity on the coordinates
does not imply the sconvexity.

The following inequality is considered the mapping connected with the

inequality (3).

Theorem 2.1

Suppose that f : = [a, b] [c, d] [0, )2 [0, ) is sconvex function

on the coordinates on . Then one has the inequalities:

 
a+b c+d
f ,
2 2
   
b d
1 c+d 1 a+b
2s2 f x, dx + f , y dy
ba a
2 dc c
2
b d
1
f (x, y) dydx (5)
(b a) (d c) a c

b
1 
b
1
1
f (x, c) dx + f (x, d) dx
2 (s + 1) b a a
ba a

d d
1 1
+ f (a, y) dy + f (b, y) dy
dc c
dc c
f (a, c) + f (b, c) + f (a, d) + f (b, d)
.
(s + 1)2
Hadamards inequality 633

Proof.

Since f : R is co-ordinated sconvex on it follows that the mapping

gx : [c, d] [0, ), gx (y) = f (x, y) is sconvex on [c, d] for all x [a, b]. Then
by sHadamards inequality (3) one has:
  d
c+d 1 gx (c) + gx (d)
2 s1
gx gx (y) dy , x [a, b] .
2 dc c
s+1

That is,
  d
c+d 1 f (x, c) + f (x, d)
2s1 f x, f (x, y) dy , x [a, b] .
2 dc c
s+1

Integrating this inequality on [a, b], we have

b   b d
2s1 c+d 1
f x, dx f (x, y) dydx (6)
ba a
2 (b a) (d c) a c

b b
1 1 1
f (x, c) dx + f (x, d) dx .
s+1 ba a
ba a

A similar arguments applied for the mapping gy : [a, b] [0, ), gy (x) =

f (x, y), we get
d   d b
2s1 a+b 1
f , y dy f (x, y) dxdy (7)
dc c
2 (d c) (b a) c a

d d
1 1 1
f (a, y) dy + f (b, y) dy .
s+1 dc c
dc c

Summing the inequalities (6) and (7), we get the second and the third inequal-
ities in (5).

Therefore, by sHadamards inequality (3), we also have:

  d  
a+b c+d 2s1 a+b
f , f , y dy (8)
2 2 dc c
2

and
  b  
a+b c+d 2s1 c+d
f , f x, dx (9)
2 2 ba a
2
634 M. Alomari and M. Darus

which give, by addition the rst inequality in (5).

Finally, by the same inequality we can also state:

b
1 f (a, c) + f (b, c)
f (x, c) dx
ba a
s+1
b
1 f (a, d) + f (b, d)
f (x, d) dx
ba a
s+1

1 
d
f (a, c) + f (a, d)
f (a, y) dy
dc c s+1

and
d
1 f (b, c) + f (b, d)
f (b, y) dy
dc c
s+1

which give, by addition the last inequality in (5).

Note 1: In (5) if s = 1 then the inequality reduced to inequality (4) .

Now, for the mapping H we have the following result(s):

Theorem 2.2

Suppose that f : R is co-ordinated sconvex on . Then for H (t, r)

one has the inequalities

(i) H is coordinated sconvex on [0, 1]2 .

(ii) The mapping H is monotonic nondecreasing on the co-ordinates.

(iii) One has the bounds

inf H (t, r) = H (0, 0) ,

(t,r)[0,1]2

and

sup H (t, r) = H (1, 1)

(t,r)[0,1]2
Hadamards inequality 635

Proof.

(i) Fix r [0, 1]. Then for all , 0 with + = 1 and t1 , t2 [0, 1],
we have:
1
H (t1 + t2 , r) =
(b a) (d c)
b d  
a+b c+d
f (t1 + t2 ) x + (1 (t1 + t2 )) , ry + (1 r) dydx
a c
2 2
b d   
1 a+b
= f t1 x + (1 t1 )
(b a) (d c) a c 2
  
a+b c+d
+ t2 x + (1 t2 ) , ry + (1 r) dydx
2 2
b d  
1 a+b c+d
= s
f t1 x + (1 t1 ) , ry + (1 r) dydx
(b a) (d c) a c 2 2
b d  
1 a+b c+d
+ s
f t2 x + (1 t2 ) , ry + (1 r) dydx
(b a) (d c) a c 2 2
= s H (t1 , r) + s H (t2 , r) .

Similarly, if t [0, 1] is xed, then for all r1 , r2 [0, 1] and , 0 with

+ = 1 we also have:

H (t, r1 + r2 ) s H (t, r1 ) + s H (t, r2 )

and the statement is proved.

(ii) Firstly, we will show that

H (t, r) H (0, r) , t, r [0, 1]2 . (10)

By sHadamards inequality (3), we have

d b  
1 1 a+b c + d
H (t, r) f tx + (1 t) dx, ry + (1 r) dy
dc c
ba a
2 2
d  
1 a+b c+d
= f , ry + (1 r) dy = H (0, r) .
dc c
2 2

t, r [0, 1]2 .
636 M. Alomari and M. Darus

Now, let 0 t1 < t2 1. By the sconvexity of the mapping H (, r) for

all r [0, 1], we have
H (t2 , r) H (t1 , r) H (t1 , r) H (0, r)
0.
t2 t1 t1
Note that, for the last inequality we have used (10).

(iii) Since f is sconvex on the coordinates, we have by Jensens inequlaity

for integrals that:
 
b d
1
1 a+b c+d
H (t, r) = f tx + (1 t) , ry + (1 r) dy dx
ba a
dc c
2 2

b d  
1 a+b 1 c+d
f tx + (1 t) , ry + (1 r) dy dx
ba a
2 dc c
2
b  
1 a+b c+d
= f tx + (1 t) , dx
ba a
2 2

b  
1 a+b c + d
f tx + (1 t) dx,
ba a
2 2
 
a+b c+d
=f ,
2 2
= H (0, 0) .
By the sconvexity of H on the co-ordinates, we have
 
b d
1
1 a+b
H (t, r) = f tx + (1 t) , y dy
ba a
dc c
2
 
d
1 a+b c+d
+ (1 r) f tx + (1 t) , dy dx
dc c
2 2

d b
1 t
1
r f (x, y) dxdy
dc c
ba a
 
b
1 a+b
+ (1 t) f , y dx dy
ba a
2
 
1  1 
d b
c+d
+ (1 r) t f x, dx
dc c ba a 2
 
a+b c+d
+ (1 t) f , dy
2 2
Hadamards inequality 637

b d d  
1 1 a+b
= rt. f (x, y) dydx + r (1 t) f , y dy
(b a) (d c) a c
dc c
2
b    
1 c+d a+b c+d
+t (1 r) f x, dx + (1 r) (1 t) f ,
ba a
2 2 2

Therefore, by (6), (7), (8) and (9) we deduce that

H (t, r) [rt + r (1 t) + t (1 r) + (1 t) (1 r)]

b d
1
f (x, y) dydx
(b a) (d c) a c
b d
1
= f (x, y) dydx
(b a) (d c) a c
= H (1, 1) .

Thus, the second bound in (iii) is proved.

Note 2: In Theorem 2.2 set s = 1 we get the result obtained in Theorem

B.

Corollary 2.1

Suppose that f : R is co-ordinated sconvex on . Dene the map-

ping h : [0, 1] R, h (t) = H (t, t). Then h is convex monotonic nondecreasing
on [0, 1] and one has the bounds:

inf h (t) = h (0)

t[0,1]

= H (0, 0) ,

and

sup h (t) = h (1)

t[0,1]

= H (1, 1) .

Proof.

Its an immediate consequence of Theorem 2.2.

638 M. Alomari and M. Darus

Comment(s):

In the next paper we will give the most main applications on these result(s)
by obtaining some Hadamardtype inequality for coordinated sconvex func-
tions in a rectangle from the plane.

Acknowledgement : The work here is supported by the Grant: UKM

GUPTMK0702107.

References
[1] H. Hudzik, L. Maligranda, Some remarks on s-convex functions, Aequa-
tiones Math., 48 (1994), 100?111.

[2] S.S. Dragomir, S. Fitzpatrick, The Hadamards inequality for s-convex

functions in the second sense, Demonstratio Math., 32 (4) (1999), 687
?696.

[3] S. S. Dragomir, On Hadamards inequality for convex functions on the

co-ordinates in a rectangle from the plane, Taiwanese Journal of Mathe-
matics, 5 (2001), 775788.

[4] S. S. Dragomir, A mapping in connection to Hadamards inequality, An

Ostro. Akad. Wiss. Math. -Natur (Wien), 128 (1991), 1720.

[5] S. S. Dragomir, Two mappings in connection to Hadamards inequality,

J. Math. Anal. Appl., 167 (1992), 4956.

Received: January 12, 2008