Professional Documents
Culture Documents
http://dx.doi.org/10.18576/amis/100507
Abstract: In this paper, we consider (s, m)-convex functions in second sense which were introduced and studied by N. Eftekhari .
We prove several Féjer-Hermite-Hadamard type integral inequalities for (s, m)-convex functions in second sense. Our results include
several new and known results as special cases. We expect that the ideas and techniques used in this paper may inspire interested readers
to explore some new applications of these newly introduced functions in various fields of pure and applied sciences.
1 Introduction
b
a+b 1
Z
f ≤ f (x) dx (1)
In recent years, theory of convex functions has received 2 (b − a) a
special attention by many researchers because of its f (a) + f (b)
importance in different fields of science like biology, ≤ . (2)
2
economy and optimization, in particular, in recent years
various generalizations and extensions of the classical In the year 1905 Leopold Féjer [8] gave a
convexity have been introduced and so the theory of generalization of the inequality (1) as the following:
inequalities has made significant contributions in various Theorem 2.If f : [a, b] ⊂ R → R is a convex function, and
areas of mathematics. g : [a, b] → R is nonnegative, integrable and symmetric
This research is dedicated to generalize some results a+b
about , then
related to convex functions, attributed to Charles Hermite 2
[10], Jaques Hadamard [9] and Lipót Fejér [8].
a+b
Z b
The Hermite-Hadamard and Fejér inequalities have f g(x)dx
2 a
been under intense investigation. Many applications and Z b
generalizations; its proofs can be found in the literature, ≤ f (x)g(x)dx
for example, [1,3,13,15,17,18,23], and their references. a
A function f : I ⊂ R → R is said to be convex function, f (a) + f (b) b
Z
if ≤ g(x)dx. (3)
2 a
f (tx + (1 − t)y) ≤ t f (x) + (1 − t) f (y), For some results which generalize, improve, and
extend the inequalities (1), see [4,5].
for all x, y ∈ I and t ∈ [0, 1]. In the year 1984, G. Toader, [20], defined
The following double inequality, which is well known m−convexity, an intermediate between usual convexity
as the Hermite-Hadamard inequality. and the starshaped property, as the following:
Theorem 1(See [9]). Let f : I ⊆ R → R be a convex Definition 1(see [20]). The function f : [0, b] → R, b > 0,
function on the interval I of real numbers and a, b ∈ I is said to be m-convex, where m ∈ [0, 1], if we have
with a < b. Then f (tx + m(1 − t)y) ≤ t f (x) + m(1 − t) f (y) (4)
∗ Corresponding author e-mail: mvivas@ucla.edu.ve
c 2016 NSP
Natural Sciences Publishing Cor.
1690 M. Vivas: Féjer type inequalities for (s, m)-convex...
for all x, y ∈ [0, b] and t ∈ [0, 1]. Definition 2(see [2]). Let 0 < s ≤ 1. A function f : [0, b] →
R, is said to be s−convex in the second sense if:
Denote by Km (b) the set of m-convex functions on
[0, b] for which f (0) ≤ 0.
f (tx + (1 − t)y) ≤ t s f (x) + (1 − t)s f (y),
Remark.For m = 1, we recapture the concept of convex
functions defined on [0, b], and m = 0 the concept of for all x, y ∈ [0, b], and t ∈ [0, 1].
starshaped functions on [0, b], i.e. f (tx) ≤ t f (x), for all S. S. Dragomir and S. Fitzpatrik (see [5]) prove the
x ∈ [0, b] and t ∈ [0, 1]. following theorem:
In the same article we can find the following results.
Theorem 5.Let f : [0, +∞) → [0, +∞) be a s−convex
Lemma 1.[21] mapping in the second sense with s ∈ (0, 1). If
a, b ∈ [0, +∞), with a < b, f ∈ L1 ([a, b]), then:
1.If f ∈ Km (b), then f it is starshaped.
2.If f is m-convex and 0 < n < m ≤ 1, then f is n-convex.
a+b
1
Z b
f (a) + f (b)
2s−1 f ≤ f (x)dx ≤ .
2 b−a a s+1
It is important to note that for m ∈ (0, 1) there are
continuous and differentiable functions that are m-convex,
but they aren’t convex in the standard sense (Ver [22]). In the year 1993 V. Mihesan (see [16]), introduced the
In [6], S.S. Dragomir and G. Toader prove the class of (s, m)− convex functions as the following:
following inequality type Hermite-Hadamard:
Definition 3([16]). The function f : [0, b] → R is said to be
Theorem 3([6]). Let f : [0, +∞) → R to be a m-convex (s, m)− convex, where s, m ∈ (0, 1], if for every x, y ∈ [a, b]
function with m ∈ (0, 1]. If 0 ≤ a < b < +∞ and and t ∈ [0, 1] we have
f ∈ L1 ([a, b]), then one has the inequality:
1 b Z
f (tx + m(1 − t)y) ≤ t s f (x) + m(1 − t s) f (y).
f (x)dx (5)
b−a a
b a
f (a) + m f
Recently, in the year 2014, combining notions of s -
(b) +
m f m f
≤ min , m . (6) convexity in the second sense of Breckner with
2 2
m-convexity, N. Eftekhari (see [7]) introduced the class of
(s, m) - convexity in the second sense.
There are some generalizations of this result, which
can be found in [11]. Definition 4([7]). A function f : [0, b] → R is said to be
Also in the year 2008, prove the following theorem of (s, m)−convex in the second sense, where (s, m) ∈ (0, 1]2
Hermite-Hadamard type for differentiable functions. if for every x, y ∈ [0, b] and t ∈ [0, 1] we have
Theorem 4([12] ). Let I be an open real interval such that
[0, +∞) ⊆ I. Let f : I → R be a differentiable on I such that f (tx + m(1 − t)y) ≤ t s f (x) + m(1 − t)s f (y).
f ′ ∈ L1 [a, b] where 0 ≤ a < b < +∞. If | f ′ |q is m−convex
on [a, b] for some fixed m ∈ (0, 1] and q ∈ [1, +∞), then Now gives way to the main objective of this paper,
Z b establish inequality of the Hermite-Hadamard-Fejér type
1
f (x)dx for functions (s, m)-convex in the second sense.
b−a a
1
b q q
′ q
| f (a)| + m| f ( )|
′
b−a m
≤ min ,
4
2 2 Some basic properties
1 Proposition 1.Let f : [0, +∞) → R be a (s, m)-convex
a
| f ′ (b)|q + m| f ′ ( )|q q function in second sense, where s, m ∈ (0, 1], and let a, b
m . be nonnegative real numbers with a < b. Then for any
2
x ∈ [a, b] there is t ∈ [0, 1] such that
f (a + b − x)
In the year 1978 , W.W. Breckner (see [2]), introduced
the definition of s− convex of real valued functions in a b
≤ t s ( f (b) + f (a)) + m(1 − t)s f +f − f (x) .
second sense: m m
c 2016 NSP
Natural Sciences Publishing Cor.
Appl. Math. Inf. Sci. 10, No. 5, 1689-1696 (2016) / www.naturalspublishing.com/Journals.asp 1691
Proposition 3.Let fn : [0, +∞) → R be is a sequence of Proof.Let x ∈ [a, b], then there is a t ∈ [0, 1] such that x =
functions. If fn is (s, m)-convex function in second sense, ta + (1 − t)b, we have
for all n ≥ k and fn (x) → f (x) ( on [0, +∞)), then f is f (x) = f (ta + (1 − t)b)
(s, m)-convex function in second sense.
b
s s
≤ t f (a) + m(1 − t) f
m
Proof.If x, y ∈ [0, +∞) and t ∈ [0, 1], we have, for n ≥ k.
b
≤ f (a) + f .
fn (tx + m(1 − t)y) ≤ t s fn (x) + m(1 − t)s fn (y), ∀n ≥ k m
s s
lim fn (tx + m(1 − t)y) ≤ lim {t fn (x) + m(1 − t) fn (y)} Thus, f is upper bounded on [a, b].
n→∞ n→∞
f (tx + m(1 − t)y) ≤ t s f (x) + m(1 − t)s f (y), Now we notice that any x ∈ [a, b] can be written as
a+b b−a
+ t for |t| ≤ (see [14]),
this completes the proof. 2 2
a+b 1 a+b 1 a+b
Proposition 4.Let f : [0, +∞) → R be a (s1 , m)−convex f = f +t + −t
function in second sense and let g : [0, +∞) → R be a 2 2 2 2 2
a+b
(s2 , m)-convex function in second sense, where − t
1 a+b m
s1 , s2 , m ∈ (0, 1]. Then f + g is (s, m)-convex function in ≤ sf +t + s f 2 .
second sense, where s = min{s1 , s2 }. 2 2 2 m
c 2016 NSP
Natural Sciences Publishing Cor.
1692 M. Vivas: Féjer type inequalities for (s, m)-convex...
Applying
the first part of this theorem in the interval Remark.Note that if we do m = 1 in the previous theorem
a b we obtain inequality of the Hermite-Hadamard-Féjer type
, , we obtain that for s-convex functions, i.e.:
m m
a+ba b Z b
f − f +f
2 m m2 f (x)g(x)dx
a
a+b
b−x s
s
− t Z b
a+b
m f (a) + f (b) x−a
≤ f − s f 2 ≤ + g(x)dx.
2 2 m 2 b−a b−a
a
Corollary 1.Under the same hypotheses of theorem 7, if
1 a+b 1
≤ f +t = f (x). g(x) = 1, we have:
2s 2 2s
Which we obtain that 1 b Z
a+b
a
b f (x) dx
2s f − 2s f +f ≤ f (x), (b − a) a
2 m m2
f (a) + f (b) m a b 1
≤ + f +f .
2 2 m m 2s
3 Main results Proof.If we take g(x) = 1 in theorem 7, we have:
The proof of the following theorem is a technique similar Z b
to the theorem 5 in [19].
f (x)dx
Theorem 7.Let f : I ⊆ [0, +∞) → [0, +∞) be a a
f (a) + f (b) b b − x s
Z
(s, m)−convex in second sense function, a, b ∈ I with
a < b, f ∈ L1 ([a, b]) and g : [a, b] → R is nonnegative, ≤ dx
2 a b−a
a+b
x−a s
Z b
integrable and symmetric about . Then m a b
2 + f +f dx.
Z b 2 m m a b−a
f (x)g(x)dx
a Given that the function ϕ (x) = xs is concave if 0 < s ≤ 1,
f (a) + f (b) b b − x s
Z then from Jensen’s inequality:
≤ g(x)dx
2 a b−a
1 bZ
x−a s
Z b
m a b f (x)dx
+ f +f g(x)dx. b−a a
2 m m a b−a s
Z b
Proof.Since f and g are real nonnegative functions, g is f (a) + f (b) 1 b−x
≤ dx
a+b 2 b−a a b−a
integrable and symmetric about , we will have that Z b s
2 m a b 1 x−a
Z b + f +f dx
f (x)g(x)dx 2 m m b−a a b−a
a
1 f (a) + f (b) m a b
= s + f +f
Z b Z b
1
= f (x)g(x)dx + f (a + b − x)g(a + b − x)dx 2 2 2 m m
2 a a
1 a b
1 b = s+1 f (a) + f (b) + m f +f .
Z
= [ f (x) + f (a + b − x)]g(x)dx 2 m m
2 a
Z b The following theorem will serve for a lower level to
1 b−x x−a
= f a +b the left inequality (3), for functions (s, m)-convex in the
2 a b−a b−a
x−a
b−x
second sense.
+f a +b g(x)dx
b−a b−a Theorem 8.Let f : I ⊆ R → [0, +∞) be is (s, m)−convex
1 b
Z
b−x s
x−a s
b in second sense function, a, b ∈ I with a < b,
≤ f (a) + m f f ∈ L1 [c, d], where c = min{a/m, a}and d = max{b, b/m}
2 a b−a b−a m
and g : [a, b] → R is nonnegative, integrable and
x−a s a b−x s
+m f + f (b) g(x)dx a+b
b−a m b−a symmetric about . Then
2
f (a) + f (b) b b − x s
Z
= g(x)dx
a+b
Z b
2 a b−a f g(x)dx
m
a
Z b
b x−a s
2 a
+ f +f g(x)dx.
1 b m b x
Z Z
2 m m a b−a
≤ s f (x)g(x)dx + s f g(x)dx.
The proof is complete. 2 a 2 a m
c 2016 NSP
Natural Sciences Publishing Cor.
Appl. Math. Inf. Sci. 10, No. 5, 1689-1696 (2016) / www.naturalspublishing.com/Journals.asp 1693
where
1 b b
Corollary 2.Under the same hypotheses of theorem 8, if M1 := f (a)g(a) + m1 m2 f g
g(x) = 1 and m = 1, we have: s1 + s2 + 1 m1 m2
s1 s2Γ (s1 )Γ (s2 )
+
1 2 +1)(s1+ s2 )Γ (s1 + s
(s + s 2)
b b
Z b · m2 f (a)g + m1 f g(a)
a+b 1 m2 m1
f (b − a) ≤ s−1 f (x)dx. and
2 2 a
1 a a
M2 := f (b)g(b) + m1 m2 f g
s1 + s2 + 1 m1 m2
In this case we obtain inequality on the side left of the s1 s2Γ (s1 )Γ (s2 )
+
theorem 5. (s1 + s2 +1)(s1+ s2 )Γ (s1 + s
2)
a a
We finally establish a Hermite-Hadamard type · m2 f (b)g + m1 f g(b) .
inequalities for the product of functions when the product m2 m1
is a function (s, m) - convex in the second sense. Proof.We have
y
f (tx + (1 − t)y) ≤ t s1 f (x) + m1 (1 − t)s1 f
Theorem 9.Let f : [0, +∞) → R be a (s, m)-convex in m1
y
second sense, where s, m ∈ (0, 1], and let a, b ∈ [0, +∞) g(tx + (1 − t)y) ≤ t s2 g(x) + m2 (1 − t)s2 g ,
m2
with a < b. Suppose that f ∈ L1 [a, b/m], and that for all t ∈ [0, 1]. f and g are nonnegative, hence
g : [a, b] → R is a nonnegative, integrable function which f (ta + (1 − t)b) · g(ta + (1 − t)b)
a+b b
is symmetric with respect to . Then ≤ t s1 +s2
f (a)g(a) + m2 t s1 (1 − t)s2 f (a)g
2 m2
Zb Zb
b
a+b x +m1 (1 − t)s1 f t s2 g(a)
2s f g(x)dx − m f g(x)dx m
2 m 1
a a b b
Zb +m1 (1 − t)s1 f m2 (1 − t)s2 g .
m1 m2
≤ f (x)g(x)dx. Integrating both sides of the above inequality over
a [0, 1], we obtain
c 2016 NSP
Natural Sciences Publishing Cor.
1694 M. Vivas: Féjer type inequalities for (s, m)-convex...
Z b Z 1
1
f (t)g(t)dt = f (ta − (1 − t)b)g(ta + (1 − t)b)dt
b−a a Z 0 Analogously we obtain
1
Z 1
b
≤ f (a)g(a) t s1 +s2 dt + m2 f (a)g t s1 (1 − t)s2 Z b
0 m2 0 1
b
Z 1 f (t)g(t)dt
+m1 f g(a) (1 − t)s1 t s2 dt b−a a
m1 0 Z
1 a a
b b 1
s1 +s2
≤ f (b)g(b) + m 1 m 2 f g
+m1 m2 f g (1 − t) dt s 1 + s2 + 1 m 1 m 2
m1 m2 0
b s 1 s2 Γ (s1 )Γ (s2 )
+
= f (a)g(a)B(s1 + s2 + 1, 1) + m2 f (a)g
m2
B(s1 + 1, s2 + 1) (s1 + s2 + 1)(s1 + s2 )Γ (s1 + s2 )
b a a
+m1 f g(a)B(s2 + 1, s1 + 1) · m2 f (b)g + m1 f g(b) ,
m1 m2 m1
b b
+m1 m2 f g B(1, s1 + s2 + 1), hence
m1 m2
Z b
1
f (t)g(t)dt ≤ min{M1 , M2 }.
b−a a
where B is the Euler Beta-function. Then,
Theorem 11.Let f : I ⊆ R → [0, +∞) be is (s, m)−convex
in second sense function, a, b ∈ I with a < b, f ∈
1
Z b L1 [c, d], where c = min{a/m, a} and d = max{b, b/m}.
f (t)g(t)dt Then
b−a a Z b
a+b m t
Γ (s1 + s2 + 1)Γ (1) 2s f − f dt
≤ f (a)g(a) 2 (b − a) a m
Γ (s1+ s2 + 2) 1 b
b Γ (s1 + 1)Γ (s2 + 1)
Z
+m2 f (a)g ≤ f (t)dt
m2 Γ (s1 + s2 + 2) b−a a
a
1 b
b Γ (s2 + 1)Γ (s1 + 1) ≤ min f (a) + m f , f (b) + m f . (7)
+m1 f g(a) s+1 m m
m1 Γ (s1 + s2 + 2)
Proof.Since f is (s, m)−convex in second sense on I, we
b Γ (1)Γ (s1 + s2 + 1)
b get
+m1 m2 f g
m1 m2 Γ (s1 + s2 + 2) s s
y
f (tx + (1 − t)y) ≤ t f (x) + m(1 − t) f . (8)
m
(s1 + s2 )Γ (s1 + s2 )
= f (a)g(a) It is easy to observe that
(s1 + s2 + 1)(s1 + s2 )Γ (s1 + s2 )
a+b ta + (1 − t)b + (1 − t)a + tb
s1Γ (s1 )s2Γ (s2 )
b f = f
+m2 f (a)g 2 2
m2 (s1 + s2 + 1)(s1 + s2 )Γ (s1 + s2 )
1 m (1 − t)a + tb
b s2Γ (s2 )s1Γ (s1 ) ≤ s f (ta + (1 − t)b) + s f
+m1 f g(a) 2 2 m
m1 (s1 + s2 + 1)(s1 + s2 )Γ (s1 + s2 )
Integrating over the interval (0, 1), we obtain
(s1 + s2 )Γ (s1 + s2 )
b b
+m1 m2 f g
a+b
Z 1 Z 1
(1 − t)a + tb
m1 m2 (s1 + s2 + 1)(s1 + s2 )Γ (s1 + s2 ) f ≤
1
f (ta + (1 − t)b)dt +
m
f dt
2 2s 0 2s 0 m
1 Z b Z b
= f (a)g(a) 1 m t
(s1 + s2 + 1) = f (t)dt + f dt
2s (b − a) a 2s (b − a) a m
s1 s2 Γ (s1 )Γ (s2 )
b
+m2 f (a)g
m2 (s1 + s2 + 1)(s1 + s2 )Γ (s1 + s2 ) We obtain
s1 s2 Γ (s2 )Γ (s1 )
Z b Z b
b a+b m t 1
+m1 f g(a) 2s f − f dt ≤ f (t)dt. (9)
m1 (s1 + s2 + 1)(s1 + s2 )Γ (s1 + s2 ) 2 (b − a) a m b−a a
b b 1 This last inequality corresponds to the first inequality
+m1 m2 f g
m1 m2 (s1 + s2 + 1) as indicated on the thesis.
On the other hand, for all t ∈ [0, 1], we have that
1 b b
= f (a)g(a) + m1 m2 f g
s1 + s2 + 1
m1 m2 b
s1 s2Γ (s1 )Γ (s2 ) f (ta + (1 − t)b) ≤ t s f (a) + m(1 − t)s f
+ m
1 2 +1)(s1+ s2 )Γ (s1 + s
(s + s 2)
b b
Integrating the above inequality over the interval (0, 1),
· m2 f (a)g + m1 f g(a) . we obtain
m2 m1
c 2016 NSP
Natural Sciences Publishing Cor.
Appl. Math. Inf. Sci. 10, No. 5, 1689-1696 (2016) / www.naturalspublishing.com/Journals.asp 1695
Now,
Z 1 Z 1 Z 1
b
Zmb Zb
f (ta + (1 − t)b)dt ≤ f (a) t s dt + m f (1 − t)s dt 1 1 1 f (a) + f (b)
0 0 m 0 f (t)dt + f (t)dt ≤ .
m + 1 mb − a b − ma s+1
1 b a ma
= f (a) + m f .
s+1 m
Remark.Note that if m = 1 in Theorem 12 we obtain the
Consequently, right side of the Hermite Hadamard inequality for
1
Z b
1
b s-convex functions.
f (x)dx ≤ f (a) + m f .
b−a a s+1 m
Zb
Using a similar procedure, we obtain that 1 f (a) + f (b)
f (t)dt ≤ . (11)
Z b b−a s+1
1 1 h a i a
f (x)dx ≤ f (b) + m f .
b−a a s+1 m We expect that the ideas and techniques used in this
Therefore we get the inequality on the right side at (7), paper may inspire interested readers to explore some new
this completes the proof of the theorem. applications of these newly introduced functions in various
fields of pure and applied sciences.
Theorem 12.Let f : [0, +∞) → [0, +∞) be is
(s, m)−convex in second sense function, a, b ∈ [0, +∞)
with a < b, s, m ∈ L1 ([0, 1]). Then
Acknowledgements
Zmb Zb
1 1 1 f (a) + f (b)
f (t)dt + f (t)dt ≤ . (10)
m + 1 mb − a b − ma s+1
a ma We want to give thanks to the library staff of Central Bank
of Venezuela (BCV) for compiling the references.
Proof.Since f is (s, m)−convex in second sense on
[0, +∞), we get
f (ta + m(1 − t)b) ≤ t s f (a) + m(1 − t)s f (b)
References
f ((1 − t)a + mtb) ≤ (1 − t)s f (a) + mt s f (b)
f (mta + (1 − t)b) ≤ mt s f (a) + (1 − t)s f (b) [1] M. Bessenyei and Zs. Páles, Characterization of convexity
f (m(1 − t)a + tb) ≤ m(1 − t)s f (a) + t s f (b) via Hadamard’s inequality, Math. Inequal. Appl. 9/1 (2006),
53-62.
[2] W. W. Breckner, Stetigkeitsaussagen für eine klasse
Adding the above four inequalities, we get verallgemeinerter konvexer funktionen in topologischen
linearen Räumen, Publ. Inst. Math. (Beograd), 23, (1978),
13-20.
f (ta + m(1 − t)b) + f ((1 − t)a + mtb)
[3] S. S. Dragomir and C. E. M. Pearce, Selected Topics
+ f (mta + (1 − t)b) + f (m(1 − t)a + tb) on Hermite-Hadamard Inequalities and Applications,
≤ ((1 − t)s + t s )(m f (b) + f (a) + f (b) + m f (a)). RGMIA Monographs, Victoria University, 2002.
(http://rgmia.vu.edu.au/monographs/).
Now, integrating over the interval (0, 1), we have [4] S. S. Dragomir, J. Pečarić and L. E.Persson , Some
Z1 Inequalities for Hadamard type, Soochow J. Math.
[ f (ta + m(1 − t)b) + f ((1 − t)a + mtb)dt 21(1995),335-341.
0
[5] S. S. Dragomir and S. Fitzpatrik, The Hadamard inequality
for s-convex functions in the second sense, Demonstration
Z1
Math. 32(4), (1999),687-696.
+ f (mta + (1 − t)b) + f (m(1 − t)a + tb)]dt
[6] S. S. Dragomir and G. Toader, Some inequalities for m-
0 convex functions, Studia Univ. Babes-Bolyai Math., 38(1),
Z1 (1993), 21-28.
≤ (m + 1)((1 − t)s + t s )( f (a) + f (b))dt. [7] N. Eftekhari, Some remarks on (s, m)−convexity in the
0 second sense, Journal of mathematical inequalities, Vol 8,
N 2 (2014), 489-495.
Where,
[8] L. Fejér. Uber die Fourierreinhen,II. Math. Naturwiss. Anz.
Zmb Zb Ungar. Akadd. Wiss. 24 (1906) 369 - 390.
2 2
f (t)dt + f (t)dt [9] J.S Hadamard. Etude sur les propiètés des fonctions entieres
mb − a b − ma
a ma et en particulier d’ une fontion considerer per Riemann, J.
Z1 Math. Pure and Appl. 58 (1893) 171- 215
≤ (m + 1)((1 − t)s + t s )( f (a) + f (b))dt. [10] Ch. Hermite, Sur deux limites d’une intégrale défine,
0
Mathesis 3, (1883), 82.
c 2016 NSP
Natural Sciences Publishing Cor.
1696 M. Vivas: Féjer type inequalities for (s, m)-convex...
c 2016 NSP
Natural Sciences Publishing Cor.