TAMKANG JOURNAL OF MATHEMATICS Volume 47, Number 3, 289-322, September 2016 doi:10.5556/j.tkjm.47.2016.1958 - - -- - + + - This paper is available online at http://journals.math.tku.edu.tw/index.php/TKJM/pages/view/onlinefirst HERMITE-HADAMARD TYPE INEQUALITIES FOR (p 1 , h 1 )-(p 2 , h 2 )-CONVEX FUNCTIONS ON THE CO-ORDINATES WENGUI YANG Abstract. In this paper, we establish some Hermite-Hadamard type inequalities for (p 1 , h1 )(p 2 , h2 )-convex function on the co-ordinates. Furthermore, some inequalities of HermiteHadamard type involving product of two convex functions on the co-ordinates are also considered. The results presented here would provide extensions of those given in earlier works. 1. Introduction The following double integral inequality f ³a +b ´ 2 ≤ 1 b−a Zb f (x)d x ≤ a f (a) + f (b) , 2 (1.1) which holds for any convex function f : [a, b] ⊂ R → R, is well known in the literature as the Hermite-Hadamard inequality, see [1, 2]. Since Hermite-Hadamard’s inequality for convex functions has been considered the most useful inequality in mathematical analysis, it has received renewed attention in recent years and a remarkable variety of refinements and generalizations have been found, we would like to refer the reader to the papers [3, 4, 5, 6, 7, 8, 9, 10, 11, 12] and the references cited therein. For example, Dragomir and Fitzpatrick [13] proved the variant of the Hermite-Hadamard inequality holds for s-convex functions in the second sense. Sarikaya et al. [14] obtained that variant of the Hadamard inequality holds for an h-convex function. Fang and Shi [15] introduced the definition of (p, h)-convex function. Definition 1.1. Let h : J → R be a nonnegative and non-zero function. We say that f : I → R is a (p, h)-convex function or that f belongs to the class g hx(h, p, I ), if f is nonnegative and 1 f ([αx p + (1 − α)y p ] p ) ≤ h(α) f (x) + h(1 − α) f (y), (1.2) Received June 21, 2015, accepted September 28, 2015. 2010 Mathematics Subject Classification. 26D15; 26A51. Key words and phrases. Hermite-Hadamard type inequalities; (p 1 , h1 )-(p 2 , h2 )-convex, co-ordinates, product of functions. 289 290 WENGUI YANG for all x, y ∈ I and α ∈ (0, 1). Similarly, if the inequality sign in (1.2) is reversed, then f is said to be a (p, h)-concave function or belong to the class g hv(h, p, I ). In [15], Fang and Shi obtained the following Hermite-Hadamard type inequalities of (p, h)convex function. Theorem 1.1. If f ∈ g hx(h, p, I ) ∩ L 1([a, b]) for a, b ∈ I with a < b, then we have Z1 Zb ³h a p + b p i p1 ´ 1 p p−1 h(t )d t . ≤ x f (x)d x ≤ [ f (a) + f (b)] f 2 bp − ap a 2h( 12 ) 0 (1.3) Furthermore, they established the following inequality of Hermite-Hadamard type involving product of two convex functions. Theorem 1.2. Suppose that f and g are functions such that f ∈ g hx(h, p, I ),∈ g hx(k, p, I ), f g ∩ L 1 ([a, b]), and hk ∩ L 1([0, 1]), with a, b ∈ I and a < b, then we have Z1 Z1 Zb p p−1 h(t )k(1 − t )d t , h(t )k(t )d t + N (a, b) x f (x)g (x)d x ≤ M (a, b) bp − bp a 0 0 where M (a, b) = f (a)g (a) + f (b)g (b) and N (a, b) = f (a)g (b) + f (b)g (a). Let us consider now a bidimensional interval ∆ = [a, b] × [c, d ] ∈ R2 with a < b and c < d . In [16], Dragomir introduced co-ordinated convex function in the following way: Definition 1.2. A mapping f : ∆ → R is said to be convex on the co-ordinates on ∆ if the inequality f (t x + (1 − t )y, r u + (1 − r )w ) ≤ t r f (x, u) + t (1 − r ) f (x, w ) + r (1 − t ) f (y, u) + (1 − t )(1 − r ) f (y, w ), holds for all t , r ∈ [0, 1] and (x, u), (y, w ) ∈ ∆. For such a mapping Dragomir [16] proved the following Hermite-Hadamard type inequalities: Theorem 1.3. Suppose that f : ∆ → R is convex on the co-ordinates on ∆. Then one has the inequalities: f Zd ³ ³ a + b c + d ´ 1 ³ 1 Zb ³ c + d ´ 1 a +b ´ ´ ≤ dx + f x, f , ,y dy 2 2 2 b−a a 2 d −c c 2 Zb Zd 1 ≤ f (x, y)d xd y (b − a)(d − c) a c Z Zb b 1 1³ 1 f (x, c)d x + f (x, d )d x ≤ 4 b−a a b−a a Zd Zd ´ 1 1 f (a, y)d y + f (b, y)d y + d −c c d −c c f (a, c) + f (a, d ) + f (b, c) + f (b, d ) . ≤ 4 HERMITE-HADAMARD TYPE INEQUALITIES 291 This idea of Dragomir inspired many researchers to extend various generalizations of convex functions on co-ordinates. In 2008, Alomari and Darus [17] defined the co-ordinated s-convexity in the second sense as follows: Definition 1.3. A mapping f : ∆ → R is said to be s-convex in the second sense on the coordinates on ∆ if the inequality f (t x + (1 − t )y, r u + (1 − r )w ) ≤ t s r s f (x, u) + t s (1 − r )s f (x, w ) + r s (1 − t )s f (y, u) + (1 − t )s (1 − r )s f (y, w ), holds for all t , r ∈ [0, 1], (x, u), (y, w ) ∈ ∆ and for some fixed s ∈ [0, 1]. For such a mapping they proved the following Hermite-Hadamard type inequalities: Theorem 1.4. Suppose that f : ∆ → R is co-ordinated s-convex on ∆. Then one has the inequalities: 4 s−1 Zd ³ ³ 1 Zb ³ c + d ´ ³a +b c +d ´ a +b ´ ´ 1 s−2 f x, f , ,y dy ≤2 dx + f 2 2 b−a a 2 d −c c 2 Zb Zd 1 ≤ f (x, y)d xd y (b − a)(d − c) a c Zb Zb 1 1 ³ 1 f (x, c)d x + f (x, d )d x ≤ 2(s + 1) b − a a b−a a Zd Zd ´ 1 1 f (a, y)d y + f (b, y)d y + d −c c d −c c f (a, c) + f (a, d ) + f (b, c) + f (b, d ) ≤ . (s + 1)2 For refinements and counterparts of other type convex functions on the co-ordinates, see [18, 19, 20, 21, 22, 23, 24, 25]. In [26], Latif and Alomari established Hadamard-type inequalities for product of two convex functions on the co-ordinates as follow: Theorem 1.5. Suppose that f , g : ∆ → [0, ∞) are convex functions on the co-ordinates on ∆. Then one has the inequalities: Zb Zd 1 f (x, y)g (x, y)d xd y (b − a)(d − c) a c 1 1 1 ≤ L(a, b, c, d ) + M (a, b, c, d ) + N (a, b, c, d ), 9 18 36 and 4f Zb Zd ³a +b c +d ´ 1 f (x, y)g (x, y)d xd y ≤ , 2 2 (b − a)(d − c) a c 292 WENGUI YANG + 7 2 5 L(a, b, c, d ) + M (a, b, c, d ) + N (a, b, c, d ), 36 36 9 where L(a, b, c, d ) = f (a, c)g (a, c) + f (b, c)g (b, c) + f (a, d )g (a, d ) + f (b, d )g (b, d ), M (a, b, c, d ) = f (a, c)g (a, d ) + f (a, d )g (a, c) + f (b, c)g (b, d ) + f (b, d )g (b, c) + f (b, c)g (a, c) + f (b, d )g (a, d ) + f (a, c)g (b, c) + f (a, d )g (b, d ), N (a, b, c, d ) = f (b, c)g (a, d ) + f (b, d )g (a, c) + f (a, c)g (b, d ) + f (a, d )g (b, c). In [27], Ödemir and Akdemir established Hermite-Hadamard-type inequalities for product of convex functions and s-convex functions of 2-variables on the co-ordinates as follow: Theorem 1.6. f : ∆ → [0, ∞) be s 1 -convex function on the co-ordinates and g : ∆ → [0, ∞) be s 2 -convex function on the co-ordinates with a < b, c < d for some fixed s 1 , s 2 ∈ (0, 1]. Then one has the inequalities: Zb Zd 1 f (x, y)g (x, y)d xd y (b − a)(d − c) a c 1 B (s 1 + 1, s 2 + 2) ≤ M (a, b, c, d ) + [B (s 1 + 1, s 2 + 2)]2 N (a, b, c, d ), L(a, b, c, d ) + (s 1 + s 2 + 1)2 s1 + s2 + 1 and let f : ∆ → [0, ∞) be convex function on the co-ordinates and g : ∆ → [0, ∞) be s-convex function on the co-ordinates with a < b, c < d for some fixed s ∈ (0, 1]. Then one has the inequality: Zb Zd ³a +b c +d ´ 1 f (x, y)g (x, y)d xd y ≤ , 2 f 2 2 (b − a)(d − c) a c 2s + 3 s 2 + 3s + 3 + L(a, b, c, d ) + M (a, b, c, d ) (s + 1)2 (s + 2)2 (s + 1)2 (s + 2)2 s 2 + 4s + 3 N (a, b, c, d ), + (s + 1)2 (s + 2)2 2s where L(a, b, c, d ), M (a, b, c, d ), N (a, b, c, d ) are defined in Theorem 1.5, and B (⋆, ∗) is Beta function. Now we first give a definition of (p, h)-convex functions on ∆. Definition 1.4. Let h : J → R be a nonnegative and non-zero function. A mapping f : ∆ → R is said to be (p, h)-convex on the co-ordinates on ∆ if the inequality 1¢ 1 ¡ f [λx p + (1 − λ)y p ] p , [λu p + (1 − λ)w p ] p ≤ h(λ) f (x, u) + h(1 − λ) f (y, w ), holds for all λ ∈ [0, 1] and (x, u), (y, w ) ∈ ∆. HERMITE-HADAMARD TYPE INEQUALITIES 293 Furthermore, we introduce the definition of (p 1 , h 1 )-(p 2 , h 2 )-convex functions on the coordinates on ∆. Definition 1.5. Let h 1 , h 2 : J → R be two nonnegative and non-zero functions. A mapping f : ∆ → R is said to be (p 1 , h 1 )-(p 2 , h 2 )-convex on the co-ordinates on ∆ if if the partial mappings f y : [a, b] → R, f y (u) = f (u, y) and f x : [c, d ] → R, f x (v) = f (x, v), are (p 1 , h 1 )-convex with respect to u and (p 2 , h 2 )-convex with respect to v, respectively, for all y ∈ [c, d ] and x ∈ [a, b]. From the above definition it follows that if f is a co-ordinated (p 1 , h 1 )-(p 2 , h 2 )-convex function, then 1 1 ¢ ¡ f [t x p 1 + (1 − t )y p 1 ] p1 , [r u p 2 + (1 − r )w p 2 ] p2 ≤ h 1 (t )h 2 (r ) f (x, u) + h 1 (t )h 2 (1 − r ) f (x, w ) + h 1 (1 − t )h 2 (r ) f (y, u) +h 1 (1 − t )h 2 (1 − r ) f (y, w ). (1.4) Similar to the proof of [16, 17], it is easily to see that every (p, h)-convex mapping f : ∆ → R is (p, h)-(p, h)-convex on the co-ordinates, but converse is not general true. Motivated by the results mentioned above, the main aim of this paper is to establish some new Hermite-Hadamard type inequalities for (p 1 , h 1 )-(p 2 , h 2 )-convex function on the co-ordinates on ∆. Furthermore, some inequalities of Hermite-Hadamard type involving product of two convex functions on the co-ordinates are also considered. The results presented here would provide extensions of those given in earlier works. 2. Main results In this section, we will give the Hermite-Hadamard type inequalities by using (p 1 , h 1 )(p 2 , h 2 )-convex functions of two variables on the co-ordinates on ∆. Theorem 2.1. Suppose that f : ∆ → R is (p 1 , h 1 )-(p 2 , h 2 )-convex function on the co-ordinates on ∆. Then one has the inequalities: ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ 1 1 2 , ¡1¢ ¡1¢ f 2 2 4h 1 2 h 2 2 Zb Zd p1p2 ≤ p x p 1 −1 y p 2 −1 f (x, y)d xd y (b 1 − a p 1 )(d p 2 − c p 2 ) a c Z1 Z1 h 2 (t )d t . h 1 (t )d t ≤ [ f (a, c) + f (a, d ) + f (b, c) + f (b, d )] 0 0 (2.1) Proof. According to (1.4) with x p 1 = sa p 1 + (1 − s)b p 1 , y p 1 = (1 − s)a p 1 + sb p 1 , u p 2 = zc p 2 + (1 − z)d p 2 , w p 2 = (1 − z)c p 2 + zd p 2 and t = r = 21 , we find that ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ 1 2 , f 2 2 294 WENGUI YANG ≤ h1 ³1´ 2 h2 ³ 1 ´³ 2 1 1 f ([sa p 1 + (1 − s)b p 1 ] p1 , [zc p 2 + (1 − z)d p 2 ] p2 ) 1 1 1 1 + f ([sa p 1 + (1 − s)b p 1 ] p1 , [(1 − z)c p 2 + zd p 2 ] p2 ) + f ([(1 − s)a p 1 + sb p 1 ] p1 , [zc p 2 + (1 − z)d p 2 ] p2 ) 1 ´ 1 + f ([(1 − s)a p 1 + sb p 1 ] p1 , [(1 − z)c p 2 + zd p 2 ] p2 ) . (2.2) Integrating (2.2) with respect to (s, z) on [0, 1] × [0, 1], we obtain f ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ 1 2 , 2 2 ³ 1 ´ ³ 1 ´³ Z1 Z1 1 1 f ([sa p 1 + (1 − s)b p 1 ] p1 , [zc p 2 + (1 − z)d p 2 ] p2 )d sd z h2 ≤ h1 2 2 0 0 Z1 Z1 1 1 + f ([sa p 1 + (1 − s)b p 1 ] p1 , [(1 − z)c p 2 + zd p 2 ] p2 )d sd z 0 0 Z1 Z1 1 1 f ([(1 − s)a p 1 + sb p 1 ] p1 , zc p 2 + (1 − z)d p 2 ] p2 )d sd z + 0 0 Z1 Z1 ´ 1 1 f ([(1 − s)a p 1 + sb p 1 ] p1 , (1 − z)c p 2 + zd p 2 ] p2 )d sd z . + 0 0 (2.3) Using the change of the variable in (2.3), we get ¡ ¢ ¡ ¢ Z Z ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ b d 4p 1 p 2 h 1 21 h 2 12 1 2 x p 1 −1 y p 2 −1 f (x, y)d xd y, f ≤ p , 2 2 (b 1 − a p 1 )(d p 2 − c p 2 ) a c which the first inequality is proved. For the proof of the second inequality in (2.1), we first note that since f is a (p 1 , h 1 )-(p 2 , h 2 )-convex function on the co-ordinates on ∆, then, by using (1.4) with x = a, y = b, u = c and w = d , it yields that 1 ¢ 1 ¡ f [t a p 1 + (1 − t )b p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 ≤ h 1 (t )h 2 (r ) f (a, c) + h 1 (t )h 2 (1 − r ) f (a, d ) + h 1 (1 − t )h 2 (r ) f (b, c) +h 1 (1 − t )h 2 (1 − r ) f (b, d ). (2.4) Similarly, we have 1 ¢ 1 ¡ f [t a p 1 + (1 − t )b p 1 ] p1 , [(1 − r )c p 2 + r d p 2 ] p2 ≤ h 1 (t )h 2 (1 − r ) f (a, c) + h 1 (t )h 2 (r ) f (a, d ) + h 1 (1 − t )h 2 (1 − r ) f (b, c) +h 1 (1 − t )h 2 (r ) f (b, d ), (2.5) 1 1 ¢ ¡ f [(1 − t )a p 1 + t b p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 ≤ h 1 (1 − t )h 2 (r ) f (a, c) + h 1 (1 − t )h 2 (1 − r ) f (a, d ) + h 1 (t )h 2 (r ) f (b, c) +h 1 (t )h 2 (1 − r ) f (b, d ), (2.6) HERMITE-HADAMARD TYPE INEQUALITIES and 295 1 1 ¢ ¡ f [(1 − t )a p 1 + t b p 1 ] p1 , [(1 − r )c p 2 + r d p 2 ] p2 ≤ h 1 (1 − t )h 2 (1 − r ) f (a, c) + h 1 (1 − t )h 2 (r ) f (a, d ) + h 1 (t )h 2 (1 − r ) f (b, c) +h 1 (t )h 2 (r ) f (b, d ). (2.7) Adding the inequalities (2.4)-(2.7), we have 1 1 ¢ 1 1 ¢ ¡ ¡ f [t a p 1 + (1 − t )b p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 + f [t a p 1 + (1 − t )b p 1 ] p1 , [(1 − r )c p 2 + r d p 2 ] p2 1 1 ¢ 1 1 ¢ ¡ ¡ + f [(1 − t )a p 1 + t b p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 + f [(1 − t )a p 1 + t b p 1 ] p1 , [(1 − r )c p 2 + r d p 2 ] p2 ≤ [h 1 (t )h 2 (r ) + h 1 (t )h 2 (1 − r ) + h 1 (1 − t )h 2 (r ) + h 1 (1 − t )h 2 (1 − r )][ f (a, c) + f (a, d ) + f (b, c) + f (b, d )]. (2.8) Integrating (2.8) with respect to (t , r ) on [0, 1] × [0, 1] and using the change of the variable, we get p1 p2 (b p 1 − a p 1 )(d p 2 − c p 2 ) Zb Zd x p 1 −1 y p 2 −1 f (x, y)d xd y a c Z1 Z1 h 2 (t )d t . h 1 (t )d t ≤ [ f (a, c) + f (a, d ) + f (b, c) + f (b, d )] 0 0 The proof is completed. By using different method, the following inequalities will be obtained. Theorem 2.2. Suppose that f : ∆ → R is (p 1 , h 1 )-(p 2 , h 2 )-convex function on the co-ordinates on ∆. Then one has the inequalities: ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ 1 1 2 , ¡1¢ ¡1¢ f 2 2 4h 1 2 h 2 2 Zb ³ h c p 2 + d p 2 i p1 ´ p1 2 p 1 −1 ≤ x f x, dx 1 p p 1 1 2 4h 2 ( 2 )(b − a ) a Zd ³h a p 1 + b p 1 i p1 ´ p2 1 p 2 −1 y f + ,y dx 2 4h 1 ( 21 )(d p 2 − d p 2 ) c Zb Zd p1 p2 x p 1 −1 y p 2 −1 f (x, y)d xd y ≤ p (b 1 − a p 1 )(d p 2 − c p 2 ) a c Zb ´ Z1 ³ Zb p1 p 1 −1 p 1 −1 ≤ h 2 (t )d t x f (x, d )d x x f (x, c)d x + 2(b p 1 − a p 1 ) a 0 a Zd ´ Z1 ³ Zd p2 p 2 −1 p 2 −1 h 1 (t )d t y f (b, y)d y y f (a, y)d y + + 2(d p 2 − c p 2 ) c 0 c Z1 Z1 h 2 (t )d t . h 1 (t )d t ≤ [ f (a, c) + f (a, d ) + f (b, c) + f (b, d )] 0 0 (2.9) 296 WENGUI YANG Proof. Since f : ∆ → R is (p 1 , h 1 )-(p 2 , h 2 )-convex function on the co-ordinates on ∆, it follows that the mapping f x : [c, d ] → R, f x (y) = f (x, y), is (p 2 , h 2 )-convex on [c, d ] for all x ∈ [a, b]. Then by using inequalities (1.3), we can write Zd Z1 ³h c p 2 + d p 2 i p1 ´ p2 1 2 p 2 −1 ≤ p fx y f x (y)d y ≤ [ f x (c) + f x (d )] h 2 (t )d t , ∀x ∈ [a, b]. 2 d 2 − c p2 c 2h 2 ( 21 ) 0 That is, Zd ³ h c p 2 + d p 2 i p1 ´ p2 2 y p 2 −1 f (x, y)d y f x, ≤ 2 d p2 − c p2 c 2h 2 ( 21 ) Z1 h 2 (t )d t , ∀x ∈ [a, b]. ≤ [ f (x, c) + f (x, d )] 1 0 Multiplying both sides of (2.10) by p 1 x p 1 −1 b p 1 −a p 1 (2.10) and integrating with respect to x over [a, b], respec- tively, we have ³ h c p 2 + d p 2 i p1 ´ 2 p 1 −1 x f x, dx 2 2h 2 ( 21 )(b p 1 − a p 1 ) a Zb Zd p1 p2 x p 1 −1 y p 2 −1 f (x, y)d xd y ≤ p (b 1 − a p 1 )(d p 2 − c p 2 ) a c Zb ´ Z1 ³ Zb p1 p 1 −1 p 1 −1 h 2 (t )d t . x f (x, d )d x ≤ p x f (x, c)d x + b 1 − a p1 a 0 a p1 Zb (2.11) A similar arguments applied for the mapping f y : [a, b] → R, f y (x) = f (x, y), we get p2 2h 1 ( 21 )(d p 2 − d p 2 ) Zd c y p 2 −1 f ³h a p 1 + b p 1 i p1 1 2 ´ ,y dx Zb Zd p1 p2 x p 1 −1 y p 2 −1 f (x, y)d xd y (b p 1 − a p 1 )(d p 2 − c p 2 ) a c Zd ´ Z1 ³ Zd p2 p 2 −1 p 2 −1 h 1 (t )d t . y f (b, y)d y y f (a, y)d y + ≤ p d 2 − c p2 c 0 c ≤ (2.12) Summing the inequalities (2.11) and (2.12), we get the second and the third inequalities in (2.9). Now, by using the first inequality in (1.3), we also have Zb ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´ 1 p1 1 2 2 p 1 −1 f , ≤ x f x, d x, (2.13) 2 2 b p1 − a p1 a 2 2h 1 ( 12 ) and Zd ³h a p 1 + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ p2 1 2 1 p 2 −1 y f , ≤ , y d y. (2.14) 2 2 d p2 − c p2 c 2 2h 2 ( 12 ) 1 f Multiplying both sides of (2.13) and (2.14) by 1 4h2 ( 12 ) obtained results, we get the first inequality in (2.9). and 1 , 4h1 ( 12 ) respectively, and adding the HERMITE-HADAMARD TYPE INEQUALITIES 297 Finally, by using the second inequality in (1.3), we can also state that Zb Z1 p1 p 1 −1 x f (x, c)d x ≤ [ f (a, c) + f (b, c)] h 1 (t )d t , b p1 − a p1 a 0 Z1 Zb p1 p 1 −1 h 1 (t )d t , x f (x, d )d x ≤ [ f (a, d ) + f (b, d )] b p1 − a p1 a 0 Zd Z1 p2 p 2 −1 y f (a, y)d y ≤ [ f (a, c) + f (a, d )] h 2 (t )d t , d p2 − c p2 c 0 and p2 p d 2 − c p2 Zd y p 2 −1 f (b, y)d y ≤ [ f (b, c) + f (b, d )] c Z1 0 h 2 (t )d t , which give, by addition, the last inequality in (2.9). The proof is completed. Remark 1. In Theorem 2.2, letting p 1 = p 2 = 1 and h 1 (t ) = h 2 (t ) = t , Theorem 2.2 reduces to Theorem 1.3. In Theorem 2.2, letting p 1 = p 2 = 1 and h 1 (t ) = h 2 (t ) = t s , Theorem 2.2 reduces to Theorem 1.4. Similarly, from the proof of Theorem 2.1, we can obtain the following theorem. Theorem 2.3. Suppose that f : ∆ → R is (p 1 , h 1 )-(p 2 , h 2 )-convex function on the co-ordinates on ∆. Then one has the inequalities: ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ Zb Zd 1 1 2 x p 1 −1 y p 2 −1 w (x, y)d xd y , ¡1¢ ¡1¢ f 2 2 4h 1 2 h 2 2 a c Zb Zd x p 1 −1 y p 2 −1 f (x, y)w (x, y)d xd y ≤ a c ≤ [ f (a, c)+ f (a, d )+ f (b, c)+ f (b, d )] ZbZd a c x p 1 −1 y p 2 −1 h 1 ³ x p 1 −a p 1 ´ ³ y p 2 −c p 2 ´ h2 p w (x, y)d xd y. b p 1 −a p 1 d 2 −c p 2 ¡£ p1 p1 ¤ 1 £ c p2 +d p2 ¤ 1 ¢ p1 p2 where w : ∆ → [0, ∞) is a symmetric function with respect to a +b , , that is, 2 2 1 ¢ 1 £ 1 £ ¤ ¤ ¤1¢ ¡£ ¤ ¡£ p p2 p 2 p2 p1 p 1 p1 p2 p 2 p2 p 1 p1 1 , r c +(1−r )d = w (1− t )a + t b , r c +(1−r )d = w t a +(1− t )b ¡£ p ¤1 £ ¤1¢ ¡£ ¤1 £ ¤1¢ p 1 p1 p2 p 2 p2 p1 p 1 p1 p2 p 2 p2 1 w t a + (1 − t )b , (1 − r )c + r d = w (1 − t )a + t b , (1 − r )c + r d . 3. Some inequalities involving product of two convex functions In this section, we will consider some inequalities of Hermite-Hadamard type involving product of two convex functions on the co-ordinates on ∆. Theorem 3.1. Suppose that f , g : ∆ → [0, ∞) are (p 1 , h 1 )-(p 2 , h 2 )-convex and (p 1 , k 1 )-(p 2 , k 2 )convex functions on the co-ordinates on ∆, respectively. Then one has the inequality: p1p2 p p (b 1 − a 1 )(d p 2 − c p 2 ) Zb Zd a c x p 1 −1 y p 2 −1 f (x, y)g (x, y)d xd y 298 WENGUI YANG Z1 h 2 (t )k 2 (t )d t h 1 (t )k 1 (t )d t 0 0 Z1 Z1 h 2 (t )k 2 (1 − t )d t h 1 (t )k 1 (t )d t +M 2 (a, b, c, d ) 0 0 Z1 Z1 +M 3 (a, b, c, d ) h 1 (t )k 1 (1 − t )d t h 2 (t )k 2 (t )d t 0 0 Z1 Z1 h 2 (t )k 2 (1 − t )d t , h 1 (t )k 1 (1 − t )d t +M 4 (a, b, c, d ) ≤ M 1 (a, b, c, d ) Z1 (3.1) 0 0 where M 1 (a, b, c, d ) = f (a, c)g (a, c) + f (b, c)g (b, c) + f (a, d )g (a, d ) + f (b, d )g (b, d ), M 2 (a, b, c, d ) = f (a, c)g (a, d ) + f (a, d )g (a, c) + f (b, c)g (b, d ) + f (b, d )g (b, c), M 3 (a, b, c, d ) = f (a, c)g (b, c) + f (a, d )g (b, d ) + f (b, c)g (a, c) + f (b, d )g (a, d ), M 4 (a, b, c, d ) = f (a, c)g (b, d ) + f (a, d )g (b, c) + f (b, c)g (a, d ) + f (b, d )g (a, c). Proof. Since f is (p 1 , h 1 )-(p 2 , h 2 )-convex on the co-ordinates and g is (p 1 , k 1 )-(p 2 , k 2 )-convex on the coordinates on ∆, it follows that 1 1 ¢ ¡ f [t a p 1 + (1 − t )b p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 ≤ h 1 (t )h 2 (r ) f (a, c) + h 1 (t )h 2 (1−r ) f (a, d ) + h 1 (1−t )h 2 (r ) f (b, c) + h 1 (1−t )h 2 (1−r ) f (b, d ), (3.2) and 1 1 ¢ ¡ g [t a p 1 + (1 − t )b p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 ≤ k 1 (t )k 2 (r )g (a, c) + k 1 (t )k 2 (1−r )g (a, d ) + k 1 (1−t )k 2 (r )g (b, c) + k 1 (1−t )k 2 (1−r )g (b, d ). (3.3) Multiplying (3.2) and (3.3) and integrating the obtained result with respect to (t , r ) on [0, 1] × [0, 1], we obtain our inequality (3.1). Theorem 3.2. Suppose that f , g : ∆ → [0, ∞) are (p 1 , h 1 )-(p 2 , h 2 )-convex and (p 1 , k 1 )-(p 2 , k 2 )convex functions on the co-ordinates on ∆, respectively. Then one has the inequality: ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ 1 2 1 2 g , , 2 2 2 2 4h 1 ( 12 )h 2 ( 12 )k 1 ( 21 )k 2 ( 21 ) Zb Zd p1 p2 x p 1 −1 y p 2 −1 f (x, y)d xd y − p (b 1 − a p 1 )(d p 2 − c p 2 ) a c ≤ Ω1 (h 1 , h 2 , k 1 , k 2 )M 1 (a, b, c, d ) + Ω2 (h 1 , h 2 , k 1 , k 2 )M 2 (a, b, c, d ) 1 f +Ω3 (h 1 , h 2 , k 1 , k 2 )M 3 (a, b, c, d ) + Ω4 (h 1 , h 2 , k 1 , k 2 )M 4 (a, b, c, d ), (3.4) where M 1 (a, b, c, d ), M 2 (a, b, c, d ), M 3 (a, b, c, d ) and M 4 (a, b, c, d ) are defined in Theorem 3.1, and Ω1 (h 1 , h 2 , k 1 , k 2 ) = Z1 0 h 1 (t )h 2 (t )d t Z1 0 k 1 (t )k 2 (1 − t )d t + Z1 0 h 1 (t )h 2 (1 − t )d t Z1 0 k 1 (t )k 2 (t )d t HERMITE-HADAMARD TYPE INEQUALITIES 299 Z1 Z1 k 1 (t )k 2 (1 − t )d t , h 1 (t )h 2 (1 − t )d t + 0 0 Z1 Z1 Z1 Z1 k 1 (t )k 2 (t )d t h 1 (t )h 2 (1 − t )d t k 1 (t )k 2 (t )d t + h 1 (t )h 2 (t )d t Ω2 (h 1 , h 2 , k 1 , k 2 ) = 0 0 0 0 Z1 Z1 + h 1 (t )h 2 (1 − t )d t k 1 (t )k 2 (1 − t )d t , 0 0 Z1 Z1 Z1 Z1 k 1 (t )k 2 (1 − t )d t h 1 (t )h 2 (t )d t k 1 (t )k 2 (t )d t + h 1 (t )h 2 (t )d t Ω3 (h 1 , h 2 , k 1 , k 2 ) = 0 0 0 0 Z1 Z1 + h 1 (t )h 2 (1 − t )d t k 1 (t )k 2 (1 − t )d t , 0 and Ω4 (h 1 , h 2 , k 1 , k 2 ) = 0 Z1 Z1 Z1 h 1 (t )h 2 (t )d t k 1 (t )k 2 (t )d t + h 1 (t )h 2 (t )d t 0 0 Z1 Z1 + h 1 (t )h 2 (1 − t )d t k 1 (t )k 2 (t )d t . 0 0 Z1 0 k 1 (t )k 2 (1 − t )d t 0 Proof. Since f is (p 1 , h 1 )-(p 2 , h 2 )-convex on the co-ordinates and g is (p 1 , k 1 )-(p 2 , k 2 )-convex on the coordinates on ∆, it follows that f y : [a, b] → [0, ∞), f y (x) = f (x, y) and f x : [c, d ] → [0, ∞), f x (y) = f (x, y) are are (p 1 , h 1 )-convex and (p 2 , h 2 )-convex on [a, b] and [c, d ], respectively, where x ∈ [a, b], y ∈ [c, d ]. Similarly, g y : [a, b] → [0, ∞), g y (x) = g (x, y) and g x : [c, d ] → [0, ∞), g x (y) = g (x, y) are are (p 1 , k 1 )-convex and (p 2 , k 2 )-convex on [a, b] and [c, d ], respectively, where x ∈ [a, b], y ∈ [c, d ]. Using Theorem 7 in [15] and multiplying both sides of the inequalities by get 1 , 2k 1 ( 12 )k 2 ( 21 ) we ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ 1 2 1 2 g , , 2 2 2 2 4h 1 ( 21 )h 2 ( 12 )k 1 ( 21 )k 2 ( 12 ) Zb ³ h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´ p1 2 2 p 1 −1 x f x, g x, dx − 2 2 2k 1 ( 12 )k 2 ( 21 )(b p 1 − a p 1 ) a · ³ h p2 1 1 c + d p 2 i p2 ´ ³ h c p 2 + d p 2 i p2 ´ 1 f a, ≤ g a, 2 2 2k 1 ( 12 )k 2 ( 21 ) ¸ Z ³ h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´ 1 2 2 h 1 (t )h 2 (1 − t )d t g b, + f b, 2 2 0 · ³ h p2 1 1 1 c + d p 2 i p2 ´ ³ h c p 2 + d p 2 i p2 ´ + g b, f a, 2 2 2k 1 ( 12 )k 2 ( 21 ) ¸ Z 1 1 ³ h c p2 + d p2 i p ´ ³ h c p2 + d p2 i p ´ 1 2 2 h 1 (t )h 2 (t )d t . (3.5) + f b, g a, 2 2 0 and ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ 1 1 2 1 2 f , g , 2 2 2 2 4h 1 ( 21 )h 2 ( 12 )k 1 ( 21 )k 2 ( 12 ) 1 f 300 WENGUI YANG ´ ³h a p 1 + b p 1 i p1 ´ 1 , y g ,y dy 2 2 2h 1 ( 21 )h 2 ( 12 )(d p 2 − c p 2 ) c · ³h p 1 1 1 1 a + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 ´ ≤ ,c g ,c f 2 2 2h 1 ( 21 )h 2 ( 12 ) ³h a p 1 + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 ´¸ Z1 1 1 k 1 (t )k 2 (1 − t )d t ,d g ,d +f 2 2 0 · ³h p 1 1 1 1 a + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 ´ + , c g ,d f 2 2 2h 1 ( 21 )h 2 ( 12 ) ³h a p 1 + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 ´¸ Z1 1 1 ,d g ,c +f k 1 (t )k 2 (t )d t . 2 2 0 Zd p2 − y p 2 −1 f ³h a p 1 + b p 1 i p1 1 (3.6) Now, by adding (3.5) and (3.6), we obtain µh p 1 1 1 ¶ µh p 1 1 ¶ a + b p 1 i p1 h c p 2 + d p 2 i p2 a 1 + b p 1 i p1 h c p 2 + d p 2 i p2 , g , 2 2 2 2 2h 1 ( 21 )h 2 ( 12 )k 1 ( 21 )k 2 ( 12 ) ¶ µ ¶ µ h p2 Zb 1 1 h c p2 + d p2 i p c + d p 2 i p2 p1 2 p 1 −1 g x, dx x f x, − 1 1 p p 1 1 2 2 2k 1 ( 2 )k 2 ( 2 )(b − a ) a Zd ³h a p 1 + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 ´ p2 1 1 p 2 −1 y f − ,y g ,y dy 1 1 p p 2 2 2 2 2h 1 ( 2 )h 2 ( 2 )(d − c ) c · ³ h p2 1 1 1 c + d p 2 i p2 ´ ³ h c p 2 + d p 2 i p2 ´ g a, ≤ f a, 2 2 2k 1 ( 12 )k 2 ( 21 ) ³ h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´¸ Z1 2 2 h 1 (t )h 2 (1 − t )d t g b, · + f b, 2 2 0 · ³ h p2 1 1 1 c + d p 2 i p2 ´ ³ h c p 2 + d p 2 i p2 ´ + g b, f a, 2 2 2k 1 ( 12 )k 2 ( 21 ) ³ h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´¸ Z1 2 2 g a, + f b, h 1 (t )h 2 (t )d t 2 2 0 · ³h p 1 1 1 a + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 ´ 1 ,c g ,c f + 2 2 2h 1 ( 12 )h 2 ( 21 ) ³h a p 1 + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 ´¸ Z1 1 1 +f ,d g ,d · k 1 (t )k 2 (1 − t )d t 2 2 0 · ³h p 1 1 1 1 a + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 ´ + , c g ,d f 2 2 2h 1 ( 12 )h 2 ( 21 ) ³h a p 1 + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 ´¸ Z1 1 1 +f ,d g ,c k 1 (t )k 2 (t )d t . (3.7) 2 2 0 1 f Applying Theorem 1.2 to each term of right hand side of the above inequality (3.7), we have Zd ³ h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´ p2 2 2 y p 2 −1 f (a, y)g (a, y)d y f a, g a, ≤ 2 2 d p2 − c p2 c 2k 1 ( 21 )k 2 ( 21 ) 1 HERMITE-HADAMARD TYPE INEQUALITIES +[ f (a, c)g (a, c) + f (a, d )g (a, d )] Z1 k 1 (t )k 2 (1 − t )d t +[ f (a, c)g (a, d ) + f (a, d )g (a, c)] Z1 k 1 (t )k 2 (t )d t , 0 0 301 (3.8) Zd ³ h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´ p2 2 2 y p 2 −1 f (b, y)g (b, y)d y g b, ≤ p f b, 1 1 2 − c p2 2 2 d 2k 1 ( 2 )k 2 ( 2 ) c Z1 +[ f (b, c)g (b, c) + f (b, d )g (b, d )] k 1 (t )k 2 (1 − t )d t + [ f (b, c)g (b, d ) 0 Z1 k 1 (t )k 2 (t )d t , (3.9) + f (b, d )g (b, c)] 1 0 Zd h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´ p2 2 2 f a, g b, ≤ y p 2 −1 f (a, y)g (b, y)d y 2 2 d p2 − c p2 c 2k 1 ( 12 )k 2 ( 21 ) Z1 k 1 (t )k 2 (1 − t )d t +[ f (a, c)g (b, c) + f (a, d )g (b, d )] 0 Z1 +[ f (a, c)g (b, d ) + f (a, d )g (b, c)] k 1 (t )k 2 (t )d t , (3.10) 1 ³ 0 Zd ³ h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´ p2 1 2 2 y p 2 −1 f (b, y)g (a, y)d y f b, g a, ≤ p 2 2 d 2 − c p2 c 2k 1 ( 12 )k 2 ( 21 ) Z1 +[ f (b, c)g (a, c) + f (b, d )g (a, d )] k 1 (t )k 2 (1 − t )d t 0 Z1 k 1 (t )k 2 (t )d t , (3.11) +[ f (b, c)g (a, d ) + f (b, d )g (a, c)] 0 ³h a p 1 + b p 1 i p1 Zb ´ ³h a p 1 + b p 1 i p1 ´ p1 1 1 x p 1 −1 f (x, c)g (x, c)d x f , c g , c ≤ 2 2 b p1 − a p1 a 2h 1 ( 21 )h 2 ( 21 ) Z1 h 1 (t )h 2 (1 − t )d t +[ f (a, c)g (a, c) + f (b, c)g (b, c)] 0 Z1 +[ f (a, c)g (b, c) + f (b, c)g (a, c)] h 1 (t )h 2 (t )d t , (3.12) 1 0 ³h a p 1 + b p 1 i p1 Zb ´ ³h a p 1 + b p 1 i p1 ´ p1 1 x p 1 −1 f (x, d )g (x, d )d x f ,d g ,d ≤ p 2 2 b 1 − a p1 a 2h 1 ( 21 )h 2 ( 21 ) Z1 +[ f (a, d )g (a, d ) + f (b, d )g (b, d )] h 1 (t )h 2 (1 − t )d t + [ f (a, d )g (b, d ) 0 Z1 h 1 (t )h 2 (t )d t , (3.13) + f (b, d )g (a, d )] 1 1 0 Zb ´ ³h a p 1 + b p 1 i p1 ´ p1 1 , c g , c ≤ x p 1 −1 f (x, c)g (x, d )d x 2 2 b p1 − a p1 a 2h 1 ( 21 )h 2 ( 21 ) Z1 h 1 (t )h 2 (1 − t )d t +[ f (a, c)g (a, d ) + f (b, c)g (b, d )] 0 Z1 +[ f (a, c)g (b, d ) + f (b, c)g (a, d )] h 1 (t )h 2 (t )d t , (3.14) 1 f ³h a p 1 + b p 1 i p1 1 0 302 WENGUI YANG and Zb ´ ³h a p 1 + b p 1 i p1 ´ p1 1 x p 1 −1 f (x, d )g (x, c)d x , c g , c ≤ 2 2 b p1 − a p1 a 2h 1 ( 21 )h 2 ( 21 ) Z1 h 1 (t )h 2 (1 − t )d t +[ f (a, d )g (a, c) + f (b, d )g (b, c)] 0 Z1 h 1 (t )h 2 (t )d t . (3.15) +[ f (a, d )g (b, c) + f (b, d )g (a, c)] 1 f ³h a p 1 + b p 1 i p1 1 0 Using the inequalities (3.8)-(3.15) in (3.7), we get ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ 1 2 1 2 g , , 2 2 2 2 2h 1 ( 21 )h 2 ( 21 )k 1 ( 12 )k 2 ( 21 ) Zb ³ h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´ p1 2 2 p 1 −1 x f x, g x, dx − 2 2 2k 1 ( 21 )k 2 ( 12 )(b p 1 − a p 1 ) a Zd ³h a p 1 + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 ´ p2 1 1 p 2 −1 y f − ,y g ,y dy 2 2 2h 1 ( 21 )h 2 ( 12 )(d p 2 − c p 2 ) c Zd ´ Z1 ³ Zd p2 p 2 −1 p 2 −1 ≤ p h 1 (t )h 2 (1 − t )d t y f (b, y)g (b, y)d y y f (a, y)g (a, y)d y + d 2 − c p2 c 0 c Z Z Z ³ d ´ 1 d p2 p 2 −1 p 2 −1 y f (a, y)g (b, y)d y + + p y f (b, y)g (a, y)d y h 1 (t )h 2 (t )d t d 2 − c p2 c c 0 Zb ´ Z1 ³ Zb p1 p 1 −1 p 1 −1 k 1 (t )k 2 (1 − t )d t x f (x, d )g (x, d )d x x f (x, c)g (x, c)d x + + p b 1 − a p1 a 0 a Zb ³ Zb ´ Z1 p1 p 1 −1 p 1 −1 + p x f (x, c)g (x, d )d x + x f (x, d )g (x, c)d x k 1 (t )k 2 (t )d t b 1 − a p1 a a 0 Z1 Z1 k 1 (t )k 2 (1 − t )d t h 1 (t )h 2 (1 − t )d t +2M 1 (a, b, c, d ) 0 0 Z1 Z1 k 1 (t )k 2 (t )d t h 1 (t )h 2 (1 − t )d t +2M 2 (a, b, c, d ) 0 0 Z1 Z1 +2M 3 (a, b, c, d ) h 1 (t )h 2 (t )d t k 1 (t )k 2 (1 − t )d t 0 0 Z1 Z1 k 2 (t )k 2 (t )d t , (3.16) h 1 (t )h 2 (t )d t +2M 4 (a, b, c, d ) 1 f 0 and 0 ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ 1 2 1 2 f , g , 1 1 1 1 2 2 2 2 2h 1 ( 2 )h 2 ( 2 )k 1 ( 2 )k 2 ( 2 ) Zb ³ h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´ p1 2 2 p 1 −1 x f x, − g x, dx 1 1 p p 1 1 2 2 2k 1 ( 2 )k 2 ( 2 )(b − a ) a Zd ³h a p 1 + b p 1 i p1 ´ ³h a p 1 + b p 1 i p1 ´ p2 1 1 p 2 −1 y f ,y g ,y dy − 2 2 2h 1 ( 21 )h 2 ( 12 )(d p 2 − c p 2 ) c Zd ´ Z1 ³ Zd p2 p 2 −1 p 2 −1 h 1 (t )h 2 (1 − t )d t y f (b, y)g (b, y)d y y f (a, y)g (a, y)d y + ≤ p d 2 − c p2 c 0 c 1 HERMITE-HADAMARD TYPE INEQUALITIES 303 Zd ´ Z1 ³ Zd p2 p 2 −1 p 2 −1 h 1 (t )h 2 (t )d t y f (b, y)g (a, y)d y y f (a, y)g (b, y)d y + d p2 − c p2 c 0 c Zb ³ Zb ´ Z1 p1 p 1 −1 p 1 −1 + p x f (x, c)g (x, c)d x + x f (x, d )g (x, d )d x k 1 (t )k 2(1 − t )d t b 1 − a p1 a a 0 Z Z Z ´ 1 ³ b b p1 p 1 −1 p 1 −1 k 1 (t )k 2(t )d t x f (x, d )g (x, c)d x x f (x, c)g (x, d )d x + + p b 1 − a p1 a 0 a Z1 Z1 +2M 1 (a, b, c, d ) h 1 (t )h 2 (1 − t )d t k 1 (t )k 2 (1 − t )d t 0 0 Z1 Z1 k 1 (t )k 2 (t )d t h 1 (t )h 2 (1 − t )d t +2M 2 (a, b, c, d ) 0 0 Z1 Z1 k 1 (t )k 2 (1 − t )d t h 1 (t )h 2 (t )d t +2M 3 (a, b, c, d ) 0 0 Z1 Z1 +2M 4 (a, b, c, d ) h 1 (t )h 2 (t )d t k 2 (t )k 2 (t )d t . (3.17) + 0 0 By applying Theorem 7 in [15] to p 1 x p 1 −1 b p 1 −a p 1 1 f 2k 1 ( 12 )k 2 ( 21 ) ³ h p p i p1 ´ ³ h p p i p1 ´ 2 2 2 2 2 2 x, c +d g x, c +d , multiplying both 2 2 and integrating over [a, b], we have Zb ³ h c p 2 + d p 2 i p1 ´ ³ h c p 2 + d p 2 i p1 ´ p1 2 2 p 1 −1 x f x, g x, dx 2 2 2k 1 ( 21 )k 2 ( 12 )(b p 1 − a p 1 ) a Zb Zd p1p2 x p 1 −1 y p 2 −1 f (x, y)d xd y − p (b 1 − a p 1 )(d p 2 − c p 2 ) a c Zb ´ Z1 ³ Zb p1 p 1 −1 p 1 −1 k 1 (t )k 2 (1 − t )d t x f (x, d )g (x, d )d x ≤ p x f (x, c)g (x, c)d x + b 1 − a p1 a 0 a Zb ´ Z1 ³ Zb p1 p 1 −1 p 1 −1 k 1 (t )k 2 (t )d t . (3.18) x f (x, d )g (x, c)d x x f (x, c)g (x, d )d x + + p b 1 − a p1 a 0 a ³h p p i p1 ´ ³h p p i p1 ´ 1 1 1 1 a 1 +b 1 1 Similarly by applying Theorem 7 in [15] to , y g a +b , y , mul1 f 1 2 2 sides by tiplying both sides by p 2 y p 2 −1 d p 2 −c p 2 p2 2h 1 ( 21 )h 2 ( 21 )(d p 2 − c p 2 ) − (b p 1 p1p2 p − a 1 )(d p 2 Zd 2h1 ( 2 )h2 ( 2 ) and integrating over [c, d ], we have y p 2 −1 f c − c p2 ) Zb Zd a ³h a p 1 + b p 1 i p1 1 2 ´ ³h a p 1 + b p 1 i p1 ´ 1 ,y g ,y dy 2 x p 1 −1 y p 2 −1 f (x, y)d xd y c Zd ´ Z1 ³ Zd p2 p 2 −1 p 2 −1 h 1 (t )h 2 (1 − t )d t y f (b, y)g (b, y)d y y f (a, y)g (a, y)d y + ≤ p d 2 − c p2 c 0 c Z Z Z ³ d ´ 1 d p2 p 2 −1 p 2 −1 + p y f (a, y)g (b, y)d y + y f (b, y)g (a, y)d y h 1 (t )h 2 (t )d t . (3.19) d 2 − c p2 c c 0 Adding (3.16)-(3.19), we get ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ 1 1 2 1 2 f , g , 2 2 2 2 4h 1 ( 12 )h 2 ( 12 )k 1 ( 21 )k 2 ( 12 ) 304 WENGUI YANG Zb Zd p1p2 x p 1 −1 y p 2 −1 f (x, y)d xd y (b p 1 − a p 1 )(d p 2 − c p 2 ) a c Zd ³ Zd ´ Z1 p2 p 2 −1 p 2 −1 ≤ p y f (a, y)g (a, y)d y + y f (b, y)g (b, y)d y h 1 (t )h 2 (1 − t )d t d 2 − c p2 c c 0 Z Z Z ´ 1 ³ d d p2 p 2 −1 p 2 −1 h 1 (t )h 2 (t )d t y f (b, y)g (a, y)d y y f (a, y)g (b, y)d y + + p d 2 − c p2 c 0 c Zb ´ Z1 ³ Zb p1 p 1 −1 p 1 −1 k 1 (t )k 2 (1 − t )d t x f (x, d )g (x, d )d x x f (x, c)g (x, c)d x + + p b 1 − a p1 a 0 a Zb ´ Z1 ³ Zb p1 p 1 −1 p 1 −1 + p k 1 (t )k 2 (t )d t x f (x, d )g (x, c)d x x f (x, c)g (x, d )d x + b 1 − a p1 a 0 a Z1 Z1 +M 1 (a, b, c, d ) h 1 (t )h 2 (1 − t )d t k 1 (t )k 2 (1 − t )d t 0 0 Z1 Z1 k 1 (t )k 2 (t )d t h 1 (t )h 2 (1 − t )d t +M 2 (a, b, c, d ) 0 0 Z1 Z1 +M 3 (a, b, c, d ) h 1 (t )h 2 (t )d t k 1 (t )k 2 (1 − t )d t 0 0 Z1 Z1 +M 4 (a, b, c, d ) h 1 (t )h 2 (t )d t k 2 (t )k 2 (t )d t . (3.20) − 0 0 Applying Theorem 1.2 to each term of right hand side of the above inequality (3.19), we have p2 d p2 − c p2 Zd y p 2 −1 f (a, y)g (a, y)d y c Z1 k 1 (t )k 2 (t )d t ≤ [ f (a, c)g (a, c) + f (a, d )g (a, d )] 0 Z1 k 1 (t )k 2 (1 − t )d t , +[ f (a, c)g (a, d ) + f (a, d )g (a, c)] 0 p2 p 2 d − c p2 Zd y p 2 −1 f (b, y)g (b, y)d y Z1 k 1 (t )k 2 (t )d t ≤ [ f (b, c)g (b, c) + f (b, d )g (b, d )] 0 Z1 k 1 (t )k 2 (1 − t )d t , +[ f (b, c)g (b, d ) + f (b, d )g (b, c)] c 0 p2 d p2 − c p2 (3.22) Zd y p 2 −1 f (a, y)g (b, y)d y Z1 ≤ [ f (a, c)g (b, c) + f (a, d )g (b, d )] k 1 (t )k 2(t )d t 0 Z1 +[ f (a, c)g (b, d ) + f (a, d )g (b, c)] k 1 (t )k 2 (1 − t )d t , c 0 p2 p d 2 − c p2 (3.21) Zd c y p 2 −1 f (b, y)g (a, y)d y (3.23) HERMITE-HADAMARD TYPE INEQUALITIES Z1 k 1 (t )k 2 (t )d t Z1 k 1 (t )k 2 (1 − t )d t , +[ f (b, c)g (a, d ) + f (b, d )g (a, c)] ≤ [ f (b, c)g (a, c) + f (b, d )g (a, d )] 0 0 p1 b p1 − a p1 0 a 0 (3.26) Zb x p 1 −1 f (x, c)g (x, d )d x Z1 ≤ [ f (a, c)g (a, d ) + f (b, c)g (b, d )] h 1 (t )h 2 (t )d t 0 Z1 h 1 (t )h 2 (1 − t )d t , +[ f (a, c)g (b, d ) + f (b, c)g (a, d )] a 0 p1 p b 1 − a p1 (3.25) Zb x p 1 −1 f (x, d )g (x, d )d x Z1 h 1 (t )h 2 (t )d t ≤ [ f (a, d )g (a, d ) + f (b, d )g (b, d )] 0 Z1 +[ f (a, d )g (b, d ) + f (b, d )g (a, d )] h 1 (t )h 2 (1 − t )d t , p1 p 1 b − a p1 (3.24) Zb x p 1 −1 f (x, c)g (x, c)d x a Z1 ≤ [ f (a, c)g (a, c) + f (b, c)g (b, c)] h 1 (t )h 2 (t )d t 0 Z1 +[ f (a, c)g (b, c) + f (b, c)g (a, c)] h 1 (t )h 2 (1 − t )d t , p1 p 1 b − a p1 305 (3.27) Zb x p 1 −1 f (x, d )g (x, c)d x a Z1 h 1 (t )h 2 (t )d t ≤ [ f (a, d )g (a, c) + f (b, d )g (b, c)] 0 Z1 +[ f (a, d )g (b, c) + f (b, d )g (a, c)] h 1 (t )h 2 (1 − t )d t . 0 (3.28) Using the inequalities (3.20)-(3.28) in (3.19), we get ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ ³h a p 1 + b p 1 i p1 h c p 2 + d p 2 i p1 ´ 1 2 1 2 g , , 2 2 2 2 4h 1 ( 21 )h 2 ( 12 )k 1 ( 21 )k 2 ( 12 ) Zb Zd p1p2 − p x p 1 −1 y p 2 −1 f (x, y)d xd y (b 1 − a p 1 )(d p 2 − c p 2 ) a c ≤ Ω1 (h 1 , h 2 , k 1 , k 2 )M 1 (a, b, c, d ) + Ω2 (h 1 , h 2 , k 1 , k 2 )M 2 (a, b, c, d ) 1 f +Ω3 (h 1 , h 2 , k 1 , k 2 )M 3 (a, b, c, d ) + Ω4 (h 1 , h 2 , k 1 , k 2 )M 4 (a, b, c, d ), which give the desired result (3.4). This completes the proof. Remark 2. In Theorems 3.1 and 3.2, letting p 1 = p 2 = 1 and h 1 (t ) = h 2 (t ) = k 1 (t ) = k 2 (t ) = t , Theorems 3.1 and 3.2 reduce to Theorem 1.5. In Theorems 3.1 and 3.2, letting p 1 = p 2 = 1, 306 WENGUI YANG and h 1 (t ) = h 2 (t ) = t s1 , k 1 (t ) = k 2 (t ) = t s2 or h 1 (t ) = h 2 (t ) = k 1 (t ) = k 2 (t ) = t s , Theorems 3.1 and 3.2 reduces to Theorem 1.6. Theorem 3.3. Suppose that f , g : ∆ → [0, ∞) are (p 1 , h 1 )-(p 2 , h 2 )-convex and (p 1 , k 1 )-(p 2 , k 2 )convex functions on the co-ordinates on ∆, respectively. Then one has the inequality: Zb Zd ³ b p1 − x p1 ´ ³ d p2 − y p2 ´ ³ p1 p2 p 1 −1 p 2 −1 x y k g (a, c) k2 p f (x, y)d xd y 1 (b p 1 − a p 1 )(d p 2 − c p 2 ) b p1 − a p1 d 2 − c p2 a c Zb Zd ³ b p1 − x p1 ´ ³ y p2 − c p2 ´ k2 p f (x, y)d xd y +g (a, d ) x p 1 −1 y p 2 −1 k 1 p b 1 − a p1 d 2 − c p2 a c Zb Zd ³ x p1 − a p1 ´ ³ d p2 − y p2 ´ x p 1 −1 y p 2 −1 k 1 p +g (b, c) k2 p f (x, y)d xd y b 1 − a p1 d 2 − c p2 a c Zb Zd ´ ³ x p1 − a p1 ´ ³ y p2 − c p2 ´ k f (x, y)d xd y +g (b, d ) x p 1 −1 y p 2 −1 k 1 p 2 b 1 − a p1 d p2 − c p2 a c Zb Zd ³ b p1 − x p1 ´ ³ d p2 − y p2 ´ ³ p1 p2 p 1 −1 p 2 −1 x y h f (a, c) h2 p g (x, y)d xd y + p 1 (b 1 − a p 1 )(d p 2 − c p 2 ) b p1 − a p1 d 2 − c p2 a c Zb Zd ³ b p1 − x p1 ´ ³ y p2 − c p2 ´ + f (a, d ) x p 1 −1 y p 2 −1 h 1 p h2 p g (x, y)d xd y b 1 − a p1 d 2 − c p2 a c Zb Zd ³ x p1 − a p1 ´ ³ d p2 − y p2 ´ h2 p g (x, y)d xd y x p 1 −1 y p 2 −1 h 1 p + f (b, c) b 1 − a p1 d 2 − c p2 a c Zb Zd ´ ³ x p1 − a p1 ´ ³ y p2 − c p2 ´ h g (x, y)d xd y + f (b, d ) x p 1 −1 y p 2 −1 h 1 p 2 b 1 − a p1 d p2 − c p2 a c Zb Zd p1p2 ≤ p x p 1 −1 y p 2 −1 f (x, y)d xd y p (b 1 − a 1 )(d p 2 − c p 2 ) a c Z1 Z1 +M 1 (a, b, c, d ) h 1 (t )k 1 (t )d t h 2 (t )k 2 (t )d t 0 0 Z1 Z1 h 2 (t )k 2 (1 − t )d t h 1 (t )k 1 (t )d t +M 2 (a, b, c, d ) 0 0 Z1 Z1 h 2 (t )k 2(t )d t h 1 (t )k 1 (1 − t )d t +M 3 (a, b, c, d ) 0 0 Z1 Z1 +M 4 (a, b, c, d ) h 1 (t )k 1 (1 − t )d t h 2 (t )k 2(1 − t )d t , 0 0 where M 1 (a, b, c, d ), M 2 (a, b, c, d ), M 3 (a, b, c, d ) and M 4 (a, b, c, d ) are defined in Theorem 3.1. Proof. Since f is (p 1 , h 1 )-(p 2 , h 2 )-convex on the co-ordinates and g is (p 1 , k 1 )-(p 2 , k 2 )-convex on the coordinates on ∆, it follows that 1 ¢ 1 ¡ f [t a p 1 + (1 − t )b p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 ≤ h 1 (t )h 2 (r ) f (a, c) + h 1 (t )h 2 (1 − r ) f (a, d ) + h 1 (1 − t )h 2 (r ) f (b, c) + h 1 (1 − t )h 2 (1 − r ) f (b, d ), (3.29) and HERMITE-HADAMARD TYPE INEQUALITIES 1 1 ¢ ¡ g [t a p 1 + (1 − t )b p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 ≤ k 1 (t )k 2 (r )g (a, c) + k 1 (t )k 2 (1 − r )g (a, d ) + k 1 (1 − t )k 2 (r )g (b, c) + k 1 (1 − t )k 2 (1 − r )g (b, d ). 307 (3.30) By (3.29)-(3.30) and using the elementary inequality, if e ≤ f and p ≤ r , then er + f p ≤ e p + f r for all e, f , p, r ∈ R, we get the following ienquality 1 ¢³ 1 ¡ f [t a p 1 + (1 − t )b p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 k 1 (t )k 2 (r )g (a, c) + k 1 (t )k 2 (1 − r )g (a, d ) ´ +k 1 (1 − t )k 2 (r )g (b, c) + k 1 (1 − t )k 2 (1 − r )g (b, d ) 1 1 ¢³ ¡ +g [t a p 1 + (1 − t )d p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 h 1 (t )h 2 (r ) f (a, c) ´ +h 1 (t )h 2 (1 − r ) f (a, d ) + h 1 (1 − t )h 2 (r ) f (b, c) + h 1 (1 − t )h 2 (1 − r ) f (b, d ) 1 ¢ ¡ 1 1 ¢ 1 ¡ ≤ f [t a p 1 + (1 − t )b p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 g [t a p 1 + (1 − t )b p 1 ] p1 , [r c p 2 + (1 − r )d p 2 ] p2 ³ ´ + h 1 (t )h 2 (r ) f (a, c) + h 1 (t )h 2 (1 − r ) f (a, d ) + h 1 (1 − t )h 2 (r ) f (b, c) + h 1 (1 − t )h 2 (1 − r ) f (b, d ) ³ × k 1 (t )k 2 (r )g (a, c) + k 1 (t )k 2 (1 − r )g (a, d ) + k 1 (1 − t )k 2 (r )g (b, c) ´ +k 1 (1 − t )k 2 (1 − r )g (b, d ) . (3.31) By integrating the above inequality (3.31) on [0, 1]×[0, 1] with respect to t , r and by taking into account the change of variables t a p 1 + (1 − t )b p 1 = x p 1 and r c p 2 + (1 − r )d p 2 = y p 2 , we obtain the desired result. Remark 3. In Theorem 3.3, letting p 1 = p 2 = 1 and h 1 (t ) = h 2 (t ) = k 1 (t ) = k 2 (t ) = t , Theorem 3.3 reduces to Theorem 10 obtained by Ödemir and Akdemir [27]. Theorem 3.4. Suppose that f , g : ∆ → [0, ∞) are (p 1 , h 1 )-(p 2 , h 2 )-convex and (p 1 , k 1 )-(p 2 , k 2 )convex functions on the co-ordinates on ∆, respectively. Then one has the inequality: Zb Zb Zd Zd Z1 Z1 p 12 p 22 x p 1 −1 y p 1 −1 u p 2 −1 w p 2 −1 4(b p 1 − a p 1 )2 (d p 2 − c p 2 )2 a a c c 0 0 1 1 ¢ ¡ × f [t x p 1 + (1 − t )y p 1 ] p1 , [r u p 2 + (1 − r )w p 2 ] p2 1 1 ¢ ¡ ×g [t x p 1 + (1 − t )y p 1 ] p1 , [r u p 2 + (1 − r )w p 2 ] p2 d t d r d xd yd ud w Z1 Z1 Zb Zd p1 p2 p 1 −1 p 2 −1 ≤ p x y f (x, y)g (x, y)d xd y h 1 (t )k 1 (t )d t h 2 (t )k 2 (t )d t (b 1 − a p 1 )(d p 2 − c p 2 ) a c 0 0 +(Θ1 + Θ3 + Θ5 )M 1 (a, b, c, d ) + (Θ1 + Θ4 + Θ5 )M 2 (a, b, c, d ) +(Θ2 + Θ3 + Θ5 )M 3 (a, b, c, d ) + (Θ2 + Θ4 + Θ5 )M 4 (a, b, c, d ), (3.32) where M 1 (a, b, c, d ), M 2 (a, b, c, d ), M 3 (a, b, c, d ) and M 4 (a, b, c, d ) are defined in Theorem 3.1, and Θ1 = Z1 0 h 2 (t )d t Z1 0 k 2 (t )d t ³ Z1 0 h 1 (t )k 1 (t )d t ´2 Z1 0 h 2 (t )k 2 (1 − t )d t , 308 WENGUI YANG Z1 h 2 (t )k 2 (1 − t )d t , h 1 (t )k 1(1 − t )d t 0 0 0 0 0 Z1 Z1 ´2 Z1 ³ Z1 h 1 (t )k 1 (1 − t )d t , h 2 (t )k 2(t )d t k 1 (t )d t h 1 (t )d t Θ3 = 0 0 0 0 Z1 Z1 Z1 Z1 Z1 Θ4 = h 1 (t )d t k 1 (t )d t h 2 (t )k 2 (t )d r h 2 (t )k 2 (1 − t )d t h 1 (t )k 1 (1 − t )d t , 0 0 0 0 0 Z1 Z1 Z1 Z1 Z1 Z1 h 2 (t )k 2 (1 − t )d t . h 1 (t )k 1 (1 − t )d t k 2 (t )d t k 1 (t )d t h 2 (t )d t h 1 (t )d t Θ5 = Θ2 = Z1 h 2 (t )d t Z1 k 2 (t )d t h 1 (t )k 1 (t )d t 0 0 0 0 Z1 Z1 0 0 Proof. Since f is (p 1 , h 1 )-(p 2 , h 2 )-convex on the co-ordinates and g is (p 1 , k 1 )-(p 2 , k 2 )-convex on the coordinates on ∆, it follows that 1 1 ¢ ¡ f [t x p 1 + (1 − t )y p 1 ] p1 , [r u p 2 + (1 − r )w p 2 ] p2 ≤ h 1 (t )h 2 (r ) f (x, u) + h 1 (t )h 2 (1 − r ) f (x, w ) + h 1(1 − t )h 2 (r ) f (y, u) +h 1 (1 − t )h 2 (1 − r ) f (y, w ), and (3.33) 1 ¢ 1 ¡ g [t x p 1 + (1 − t )y p 1 ] p1 , [r u p 2 + (1 − r )w p 2 ] p2 ≤ k 1 (t )k 2 (r )g (x, u) + k 1 (t )k 2 (1 − r )g (x, w ) + k 1 (1 − t )k 2 (r )g (y, u) +k 1 (1 − t )k 2 (1 − r )g (y, w ), (3.34) for t , r ∈ [0, 1], x, y ∈ [a, b] and u, w ∈ [c, d ]. Because f and g are nonnegative, from (3.33) and (3.34), we get the inequality 1 ¢ ¡ 1 1 ¢ 1 ¡ f [t x p 1 + (1 − t )y p 1 ] p1 , [r u p 2 + (1 − r )w p 2 ] p2 g [t x p 1 + (1 − t )y p 1 ] p1 , [r u p 2 + (1 − r )w p 2 ] p2 ³ ´ ≤ h 1 (t )h 2 (r ) f (x, u) + h 1 (t )h 2 (1−r ) f (x, w ) + h 1 (1−t )h 2 (r ) f (y, u) + h 1(1−t )h 2 (1−r ) f (y, w ) ³ ´ × k 1 (t )k 2 (r )g (x, u) + k 1 (t )k 2 (1−r )g (x, w ) + k 1 (1−t )k 2 (r )g (y, u) + k 1 (1−t )k 2 (1−r )g (y, w ) = h 1 (t )h 2 (r )k 1 (t )k 2 (r ) f (x, u)g (x, u) + h 1 (t )h 2 (1 − r )k 1 (t )k 2 (1 − r ) f (x, w )g (x, w ) +h 1 (1−t )h 2 (r )k 1 (1−t )k 2 (r ) f (y, u)g (y, u) + h 1 (1−t )h 2 (1−r )k 1 (1−t )k 2 (1−r ) f (y, w )g (y, w ) +h 1 (t )h 2 (1 − r )k 1 (t )k 2 (r ) f (x, w )g (x, u) + h 1 (t )h 2 (r )k 1 (t )k 2 (1 − r ) f (x, u)g (x, w ) +h 1 (1−t )h 2 (1−r )k 1 (1−t )k 2 (r ) f (y, w )g (y, u) + h 1(1−t )h 2 (r )k 1 (1−t )k 2 (1−r ) f (y, u)g (y, w +h 1 (t )h 2 (r )k 1 (1 − t )k 2 (r ) f (x, u)g (y, u) + h 1 (1 − t )h 2 (r )k 1 (t )k 2 (r ) f (y, u)g (x, u) +h 1 (1−t )h 2 (1−r )k 1 (t )k 2 (1−r ) f (y, w )g (x, w ) + h 1(t )h 2 (1−r )k 1 (1−t )k 2 (1−r ) f (x, w )g (y, w ) +h 1 (1 − t )h 2 (1 − r )k 1 (t )k 2 (r ) f (y, w )g (x, u) + h 1(1 − t )h 2 (r )k 1 (t )k 2 (1 − r ) f (y, u)g (x, w ) +h 1 (t )h 2 (1−r )k 1 (1−t )k 2 (r ) f (x, w )g (y, u) + h 1(t )h 2 (r )k 1 (1−t )k 2 (1−r ) f (x, u)g (y, w ). (3.35) Multiplying both sides of (3.35) by [0, 1]2 , we have p 12 p 22 (b p 1 − a p 1 )2 (d p 2 − c p 2 )2 p 12 p 22 x p 1 −1 y p 1 −1 u p 2 −1 w p 2 −1 (b p 1 −a p 1 )2 (d p 2 −c p 2 )2 Zb Zb Zd Zd Z1 Z1 a a c c 0 0 and integrating over [a, b]2 ×[c, d ]2 × x p 1 −1 y p 1 −1 u p 2 −1 w p 2 −1 HERMITE-HADAMARD TYPE INEQUALITIES 309 1 1 ¢ ¡ × f [t x p 1 + (1 − t )y p 1 ] p1 , [r u p 2 + (1 − r )w p 2 ] p2 1 1 ¢ ¡ ×g [t x p 1 + (1 − t )y p 1 ] p1 , [r u p 2 + (1 − r )w p 2 ] p2 d t d r d xd yd ud w Zb Zd Z1 Z1 4p 1 p 2 p 1 −1 p 2 −1 x y f (x, y)g (x, y)d xd y h (t )k (t )d t ≤ p h 2 (t )k 2 (t )d t 1 1 (b 1 − a p 1 )(d p 2 − c p 2 ) a c 0 0 Zb Zd Zd 4p 1 p 22 x p 1 −1 u p 2 −1 w p 2 −1 f (x, w )g (x, u)d xd ud w + p (b 1 − a p 1 )(d p 2 − c p 2 )2 a c c Z1 Z1 h 1 (t )k 1 (t )d t h 2 (t )k 2 (1 − t )d t 0 + 0 4p 12 p 2 Zb Zb Zd x p 1 −1 y p 1 −1 u p 2 −1 f (x, u)g (y, u)d xd yd u (b p 1 − a p 1 )2 (d p 2 − c p 2 ) a a c Z1 Z1 × h 1 (t )k 1 (1 − t )d t h 2 (t )k 2(t )d t 0 + 0 4p 12 p 22 Zb Zd x p 1 −1 w p 2 −1 f (x, w )d xd w (b p 1 − a p 1 )2 (d p 2 − c p 2 )2 a c Z1 Z1 × h 1 (t )k 1 (1 − t )d t h 2 (t )k 2(1 − t )d t . 0 Zb Zd a y p 1 −1 u p 2 −1 g (y, u)d yd u c (3.36) 0 Applying Theorems 1.1 and 1.2 to second and third term of right hand side of the inequality (3.36), we have p 1 p 22 Zb Zd Zd x p 1 −1 u p 2 −1 w p 2 −1 f (x, w )g (x, u)d xd ud w (b p 1 − a p 1 )(d p 2 − c p 2 )2 a c c Zd Zd Zb ³ ´ p 22 p1 p 2 −1 p 2 −1 p 1 −1 = p u w x f (x, w )g (x, u)d x d ud w (d 2 − c p 2 )2 c c b p1 − a p1 a Z1 Zd Zd ³ p2 h 1 (t )k 1 (t )d t u p 2 −1 w p 2 −1 [ f (a, w )g (a, u) + f (b, w )g (b, u)] ≤ p 2p 2 (d 2 − c 2 ) c c 0 Z1 ´ +[ f (a, w )g (b, u) + f (b, w )g (a, u)] h 1 (t )k 1 (1 − t )d t d ud w 0 ≤ p 22 ³h Zd w p 2 −1 f (a, w )d w Zd u p 2 −1 g (a, u)d u (d p 2 − c p 2 )2 c c Zd Zd i Z1 h 1 (t )k 1 (t )d t u p 2 −1 g (b, u)d u w p 2 −1 f (b, w )d w + 0 c c Zd h Zd p 2 −1 + w f (a, w )d w u p 2 −1 g (b, u)d u c c Zd Zd i Z1 ´ h 1 (t )k 1 (1 − t )d t g (a, u)d u 0 c c Z1 Z1 Z1 ≤ [ f (a, c) + f (a, d )][g (a, c) + g (a, d )] h 2 (t )d t k 2 (t )d t h 1 (t )k 1 (t )d t 0 0 0 Z1 Z1 Z1 h 1 (t )k 1 (t )d t k 2 (t )d t h 2 (t )d t +[ f (b, c) + f (b, d )][g (b, c) + g (b, d )] + w p 2 −1 f (b, w )d w u p 2 −1 0 0 0 310 WENGUI YANG +[ f (a, c) + f (a, d )][g (b, c) + g (b, d )] Z1 0 h 2 (t )d t Z1 0 k 2 (t )d t Z1 0 h 1 (t )k 1 (1 − t )d t Z1 Z1 Z1 h 1 (t )k 1 (1 − t )d t k 2 (t )d t h 2 (t )d t +[ f (b, c) + f (b, d )][g (a, c) + g (a, d )] 0 0 0 Z1 Z1 Z1 = [M 1(a, b, c, d ) + M 2 (a, b, c, d )] h 2 (t )d t k 2 (t )d t h 1 (t )k 1(t )d t 0 0 0 Z1 Z1 Z1 h 1 (t )k 1 (1 − t )d t , k 2 (t )d t h 2 (t )d t +[M 3 (a, b, c, d ) + M 4(a, b, c, d )] and p 12 p 2 0 0 0 (3.37) Zb Zb Zd x p 1 −1 y p 1 −1 u p 2 −1 f (x, u)g (y, u)d xd yd u (b p 1 − a p 1 )2 (d p 2 − c p 2 ) a a c Zb Zb Zd ³ ´ p 12 p2 p 2 −1 p 1 −1 p 1 −1 = p u f (x, u)g (y, u)d u d xd y x y (b 1 − a p 1 )2 a a d p2 − c p2 c Zb Zb Z1 ³ p2 ≤ p 1 p 2 x p 1 −1 y p 1 −1 [ f (x, c)g (y, c) + f (x, d )g (y, d )] h 2 (t )k 2 (t )d t (b 1 − a 1 ) a a 0 Z1 ´ h 2 (t )k 2 (1 − t )d t d xd y +[ f (x, c)g (y, d ) + f (x, d )g (y, c)] 0 ≤ p 12 ³h Zb Zb p 1 −1 x f (x, c)d x y p 1 −1 g (y, c)d y (b p 1 − a p 1 )2 a a Zb Zb i Z1 + x p 1 −1 f (x, d )d x y p 1 −1 g (y, d )d y h 2 (t )k 2 (t )d t a 0 a Zb h Zb p 1 −1 x f (x, c)d x + y p 1 −1 g (y, d )d y a a Zb Zb i Z1 ´ g (y, c)d t h 2 (t )k 2 (1 − t )d t a a 0 Z1 Z1 Z1 h 2 (t )k 2 (t )d t k 1 (t )d t h 1 (t )d t ≤ [ f (a, c) + f (b, c)][g (a, c) + g (b, c)] 0 0 0 Z1 Z1 Z1 +[ f (a, d ) + f (b, d )][g (a, d ) + g (b, d )] h 1 (t )d t k 1 (t )d t h 2 (t )k 2 (t )d t 0 0 0 Z1 Z1 Z1 h 2 (t )k 2 (1 − t )d t k 1 (t )d t h 1 (t )d t +[ f (a, c) + f (b, c)][g (a, d ) + g (b, d )] 0 0 0 Z1 Z1 Z1 h 2 (t )k 2 (1 − t )d t k 1 (t )d t h 1 (t )d t +[ f (a, d ) + f (b, d )][g (a, c) + g (b, c)] 0 0 0 Z1 Z1 Z1 = [M 1(a, b, c, d ) + M 3 (a, b, c, d )] h 1 (t )d t k 1 (t )d t h 2 (t )k 2(t )d t 0 0 0 Z1 Z1 Z1 h 2 (t )k 2 (1 − t )d t . k 1 (t )d t h 1 (t )d t +[M 2 (a, b, c, d ) + M 4(a, b, c, d )] + x p 1 −1 f (x, d )d x y p 1 −1 0 0 0 (3.38) Applying Theorem 2.1 to fourth term of right hand side of the inequality (3.36), we have p 12 p 22 (b p 1 − a p 1 )2 (d p 2 − c p 2 )2 Zb Zd a c x p 1 −1 w p 2 −1 f (x, w )d xd w Zb Zd a c y p 1 −1 u p 2 −1 g (y, u)d yd u HERMITE-HADAMARD TYPE INEQUALITIES 311 Z1 Z1 ´ ³ h 2 (t )d t h 1 (t )d t ≤ [ f (a, c) + f (a, d ) + f (b, c) + f (b, d )] 0 0 Z1 Z1 ´ ³ k 2 (t )d t k 1 (t )d t · [g (a, c) + g (a, d ) + g (b, c) + g (b, d )] 0 0 = [M 1 (a, b, c, d ) + M 2(a, b, c, d ) + M 3 (a, b, c, d ) + M 4(a, b, c, d )] Z1 Z1 Z1 Z1 k 2 (t )d t . k 1 (t )d t h 2 (t )d t h 1 (t )d t × 0 0 0 0 Using the inequalities (3.37)-(3.39) in (3.36), we get the desired result (3.32). (3.39) Remark 4. Suppose that f , g : ∆ → [0, ∞) are convex functions on the co-ordinates on ∆. Then one has the inequality: Zb Zb Zd Zd Z1 Z1 9 f (t x + (1 − t )y, r u + (1 − r )w )g (t x + (1 − t )y, 4(b − a)2 (d − c)2 a a c c 0 0 r u + (1 − r )w )d t d r d xd yd ud w Zb Zd 19 5 11 1 f (x, y)g (x, y)d xd y + L(a, b, c, d )+ M (a, b, c, d )+ N (a, b, c, d ), ≤ (b − a)(d − c) a c 192 64 192 where L(a, b, c, d ), M (a, b, c, d ) and N (a, b, c, d ) are defined in Theorem 1.5. Remark 5. Suppose that f , g : ∆ → [0, ∞) are s-convex functions on the co-ordinates on ∆. Then one has the inequality: Zb Zb Zd Zd Z1 Z1 (1 + 2s)2 f (t x + (1 − t )y, r u + (1 − r )w ) 4(b − a)2 (d − c)2 a a c c 0 0 g (t x + (1 − t )y, r u + (1 − r )w )d t d r d xd yd ud w Zb Zd 1 ≤ f (x, y)g (x, y)d xd y (b − a)(d − c) a c ³ (1 + 2s)2 [Γ(1 + s)]2 + 2(1 + s)2 Γ(2 + 2s) +[Γ(1 + s)]2 L(a, b, c, d ) (1 + s)4 [Γ(2 + 2s)]2 (2 + 8s + 9s 2 + 2s 3 )[Γ(1 + s)]2 + (1 + s)2 Γ(2 + 2s) M (a, b, c, d ) + (1 + s)4 [Γ(2 + 2s)]2 ´ (3 + 6s + 2s 2 )[Γ(1 + s)]2 + N (a, b, c, d ) , (1 + s)4 [Γ(2 + 2s)]2 where L(a, b, c, d ), M (a, b, c, d ) and N (a, b, c, d ) are defined in Theorem 1.5. Remark 6. Suppose that f , g : ∆ → [0, ∞) are s 1 -convex and s 2 -convex functions on the coordinates on ∆, respectively. Then one has the inequality: (1 + s 1 + s 2 )2 4(b − a)2 (d − c)2 Zb Zb Zd Zd Z1 Z1 a a c c 0 0 f (t x + (1−t )y, r u + (1−r )w )g (t x + (1−t )y, r u + (1−r )w ) d t d r d xd yd ud w ≤ 1 (b − a)(d − c) Zb Zd a c f (x, y)g (x, y)d xd y + Ξ1L(a, b, c, d ) + Ξ2 M (a, b, c, d ) + Ξ3 N (a, b, c, d ), 312 WENGUI YANG where L(a, b, c, d ), M (a, b, c, d ) and N (a, b, c, d ) are defined in Theorem 1.5, and ¡ ¢ Γ(1 + s 1 )Γ(1 + s 2 ) (1 + s 1 + s 2 )2 Γ(1 + s 1 )Γ(1 + s 2 ) + 2(1 + s 1 )(1 + s 2 )Γ(2 + s 1 + s 2 ) Ξ1 = , (1 + s 1 )2 (1 + s 2 )2 [Γ(2 + s 1 + s 2 )]2 (2(1 + s 2 )2 + s 12 (2 + s 2 ) + s 1 (4 + 5s 2 + s 22 ))[Γ(1 + s 1 )]2 [Γ(1 + s 2 )]2 Ξ2 = (1 + s 1 )2 (1 + s 2 )2 [Γ(2 + s 1 + s 2 )]2 (1 + s 1 )(1 + s 2 )Γ(1 + s 1 )Γ(1 + s 2 )Γ(2 + s 1 + s 2 ) , + (1 + s 1 )2 (1 + s 2 )2 [Γ(2 + s 1 + s 2 )]2 (1 + s 1 + s 2 )(3 + 3s 1 + 3s 2 + 2s 1 s 2 )[Γ(1 + s 1 )]2 [Γ(1 + s 2 )]2 Ξ3 = . (1 + s 1 )2 (1 + s 2 )2 [Γ(2 + s 1 + s 2 )]2 Theorem 3.5. Suppose that f , g : ∆ → [0, ∞) are (p 1 , h 1 )-(p 2 , h 2 )-convex and (p 1 , k 1 )-(p 2 , k 2 )convex functions on the co-ordinates on ∆, respectively. Then one has the inequality: p1 p2 p p (b 1 − a 1 )(d p 2 − c p 2 ) Zb Zd Z1 Z1 a c 0 0 x p 1 −1 u p 2 −1 f ³£ t x p 1 + (1 − t ) a p 1 + b p 1 ¤ p1 1 , 2 c p 2 + d p 2 ¤ p1 ´ 2 2 ³£ a p 1 + b p 1 ¤ p1 £ p 2 c p 2 + d p 2 ¤ p1 ´ 1 , ru 2 d t d r d xd u ×g t x p 1 + (1 − t ) + (1 − r ) 2 2 Zb Zd Z1 Z1 p1p2 p 1 −1 p 2 −1 x u f (x, u)g (x, u)d xd u ≤ p h 1 (t )k 1 (t )d t h 2 (t )k 2 (t )d t (b 1 − a p 1 )(d p 2 − c p 2 ) a c 0 0 +(Λ1 + Λ3 + Λ5 )M 1 (a, b, c, d ) + (Λ1 + Λ4 + Λ5 )M 2 (a, b, c, d ) £ r u p 2 + (1 − r ) +(Λ2 + Λ3 + Λ5 )M 3 (a, b, c, d ) + (Λ2 + Λ4 + Λ5 )M 4 (a, b, c, d ), (3.40) where M 1 (a, b, c, d ), M 2 (a, b, c, d ), M 3 (a, b, c, d ) and M 4 (a, b, c, d ) are defined in Theorem 3.1, and Z Z ³ ´ ¡ 1 1 1 1 ¢ 1 1 Λ1 = 2 2h 2 ( )k 2 ( ) h 2 (t )k 2(1 − t )d t h 2 (t )k 2 (t )d t + h 2 ( ) + k 2 ( ) 2 2 0 2 2 0 Z1 Z1 ´2 ³ Z1 h 1 (t )k 1 (t )d t , k 2 (t )d t d t h 2 (t ) 0 0 0 Z Z ³ ´ Z1 ¡ 1 1 1 1 1 ¢ 1 Λ2 = 2 2h 2 ( )k 2 ( ) h 2 (t )d t h 2 (t )k 2(1 − t )d t h 2 (t )k 2 (t )d t + h 2 ( ) + k 2 ( ) 2 2 0 2 2 0 0 Z1 Z1 Z1 k 2 (t )d t h 1 (t )k 1 (t )d t h 1 (t )k 1 (1 − t )d t , 0 0 0 Z Z ´ ³ ¡ 1 1 1 1 ¢ 1 1 h 1 (t )k 1(1 − t )d t h 1 (t )k 1 (t )d t + h 1 ( ) + k 1 ( ) Λ3 = 2 2h 1 ( )k 1 ( ) 2 2 0 2 2 0 Z1 Z1 ´2 ³ Z1 h 2 (t )k 2 (t )d t , h 1 (t )d t k 1 (t )d t 0 0 0 Z Z ³ ´ ¡ 1 1 1 1 ¢ 1 1 Λ4 = 2 2h 1 ( )k 1 ( ) h 1 (t )k 1(1 − t )d t h 1 (t )k 1 (t )d t + h 1 ( ) + k 1 ( ) 2 2 0 2 2 0 Z1 Z1 Z1 Z1 h 2 (t )k 2 (1 − t )d t , h 2 (t )k 2 (t )d t k 1 (t )d t h 1 (t )d t 0 0 0 0 HERMITE-HADAMARD TYPE INEQUALITIES 313 Z1 Z1 ³ 1 1 1 1 h 2 (t )k 2 (t )d t Λ5 = 4 4h 1 ( )h 2 ( )k 1 ( )k 2 ( ) h 1 (t )k 1 (t )d t 2 2 2 2 0 0 Z1 Z 1 1 ¡ 1 1 ¢ 1 +2h 1 ( )k 1 ( ) h 2 ( ) + k 2 ( ) h 2 (t )k 2 (1 − t )d t h 1 (t )k 1 (t )d t 2 2 2 2 0 0 Z Z1 1 ¡ 1 1 ¢ 1 1 h 1 (t )k 1 (1 − t )d t +2h 2 ( )k 2 ( ) h 1 ( ) + k 1 ( ) h 2 (t )k 2 (t )d t 2 2 2 2 0 0 Z ¡ 1 1 1 1 1 1 1 ¢ 1 1 h 1 (t )k 1 (1 − t )d t + h 1 ( )k 2 ( ) + h 2 ( )k 1 ( ) + h 1 ( )h 2 ( ) + k 1 ( )k 2 ( ) 2 2 2 2 2 2 2 2 0 Z1 Z Z Z Z ´ 1 1 1 1 h 2 (t )k 2 (1 − t )d t h 1 (t )d t h 2 (t )d t k 1 (t )d t k 2 (t )d t . 0 0 0 0 0 Proof. Since f is (p 1 , h 1 )-(p 2 , h 2 )-convex on the co-ordinates and g is (p 1 , k 1 )-(p 2 , k 2 )-convex on the coordinates on ∆, it follows that ³£ c p 2 + d p 2 ¤ p1 ´ a p 1 + b p 1 ¤ p1 £ p 2 1 , ru 2 + (1 − r ) f t x p 1 + (1 − t ) 2 2 ¡ £ c p2 + d p2 ¤ 1 ¢ p2 ≤ h 1 (t )h 2 (r ) f (x, u) + h 1 (t )h 2 (1 − r ) f x, 2 ¡£ a p 1 +b p 1 ¤ 1 £ c p 2 +d p 2 ¤ 1 ¢ ¡£ a p 1 + b p 1 ¤ 1 ¢ p1 p1 p2 , u + h 1 (1−t )h 2 (1−r ) f , ,(3.41) +h 1 (1 − t )h 2 (r ) f 2 2 2 and ³£ a p 1 + b p 1 ¤ p1 £ p 2 c p 2 + d p 2 ¤ p1 ´ 1 , ru 2 g t x p 1 + (1 − t ) + (1 − r ) 2 2 ¡ £ c p2 + d p2 ¤ 1 ¢ p2 ≤ k 1 (t )k 2 (r ) f (x, u) + k 1 (t )k 2 (1 − r ) f x, 2 ¡£ a p 1 + b p 1 ¤ 1 ¢ ¡£ a p 1 +b p 1 ¤ 1 £ c p 2 +d p 2 ¤ 1 ¢ p1 p1 p2 +k 1 (1 − t )k 2 (r ) f , u + k 1 (1−t )k 2 (1−r ) f , , (3.42) 2 2 2 for t , r ∈ [0, 1], x ∈ [a, b] and u ∈ [c, d ]. Because f and g are nonnegative, from (3.41) and (3.42), we get the inequality ³£ c p 2 + d p 2 ¤ p1 ´ a p 1 + b p 1 ¤ p1 £ p 2 1 , ru 2 + (1 − r ) f t x p 1 + (1 − t ) 2 2 ³£ a p 1 + b p 1 ¤ p1 £ p 2 c p 2 + d p 2 ¤ p1 ´ 1 , ru 2 ×g t x p 1 + (1 − t ) + (1 − r ) 2 2 ³ ¡ £ c p2 + d p2 ¤ 1 ¢ ¡£ a p 1 + b p 1 ¤ 1 ¢ p2 p1 ≤ h 1 (t )h 2 (r ) f (x, u) + h 1 (t )h 2 (1 − r ) f x, + h 1 (1 − t )h 2 (r ) f ,u 2 2 ¡£ a p 1 + b p 1 ¤ 1 £ c p 2 + d p 2 ¤ 1 ¢´ p1 p2 , +h 1 (1 − t )h 2 (1 − r ) f 2 2 ³ ¡ £ c p2 + d p2 ¤ 1 ¢ p2 k 1 (t )k 2 (r ) f (x, u) + k 1 (t )k 2 (1 − r ) f x, 2 ¡£ a p 1 + b p 1 ¤ 1 ¢ ¡£ a p 1 + b p 1 ¤ 1 £ c p 2 + d p 2 ¤ 1 ¢´ p1 p1 p2 +k 1 (1 − t )k 2 (r ) f , u + k 1 (1 − t )k 2 (1 − r ) f , 2 2 2 £ c p2 + d p2 ¤ 1 p2 ) = h 1 (t )h 2 (r )k 1 (t )k 2 (r ) f (x, u)g (x, u) + h 1 (t )h 2 (1 − r )k 1 (t )k 2 (1 − r ) f (x, 2 £ a p1 + b p1 ¤ 1 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 p2 p1 p1 ) + h 1 (1 − t )h 2 (r )k 1 (1 − t )k 2 (r ) f ( , u)g ( , u) g (x, 2 2 2 314 WENGUI YANG £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p2 +h 1 (1 − t )h 2 (1 − r )k 1 (1 − t )k 2 (1 − r ) f ( , ) 2 2 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p2 p2 g( , ) + h 1 (t )h 2 (1 − r )k 1 (t )k 2 (r ) f (x, )g (x, u) 2 2 2 p p £c 2 +d 2 ¤ 1 p2 ) +h 1 (t )h 2 (r )k 1 (t )k 2 (1 − r ) f (x, u)g (x, 2 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 £ a p1 + b p1 ¤ 1 p1 p2 p1 +h 1 (1 − t )h 2 (1 − r )k 1 (1 − t )k 2 (r ) f ( , )g ( , u) 2 2 2 £ a p1 + b p1 ¤ 1 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p1 p2 +h 1 (1 − t )h 2 (r )k 1 (1 − t )k 2 (1 − r ) f ( , u)g ( , ) 2 2 2 £ a p1 + b p1 ¤ 1 p1 , u) +h 1 (t )h 2 (r )k 1 (1 − t )k 2 (r ) f (x, u)g ( 2 £ a p1 + b p1 ¤ 1 p1 +h 1 (1 − t )h 2 (r )k 1 (t )k 2(r ) f ( , u)g (x, u) 2 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p2 p2 +h 1 (1 − t )h 2 (1 − r )k 1 (t )k 2 (1 − r ) f ( , )g (x, ) 2 2 2 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 £ c p2 + d p2 ¤ 1 p2 p1 p2 )g ( , ) +h 1 (t )h 2 (1 − r )k 1 (1 − t )k 2 (1 − r ) f (x, 2 2 2 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p2 +h 1 (1 − t )h 2 (1 − r )k 1 (t )k 2 (r ) f ( , )g (x, u) 2 2 p p p £a 1 +b 1 ¤ 1 £ c 2 + d p2 ¤ 1 p1 p2 +h 1 (1 − t )h 2 (r )k 1 (t )k 2(1 − r ) f ( , u)g (x, ) 2 2 £ c p2 + d p2 ¤ 1 £ a p1 + b p1 ¤ 1 p2 p1 +h 1 (t )h 2 (1 − r )k 1 (1 − t )k 2 (r ) f (x, )g ( , u) 2 2 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p2 , ). (3.43) +h 1 (t )h 2 (r )k 1 (1 − t )k 2 (1 − r ) f (x, u)g ( 2 2 Multiplying both sides of (3.43) by p 1 p 2 x p 1 −1 u p 2 −1 (b p 1 −a p 1 )(d p 2 −c p 2 ) and integrating over [a, b] × [c, d ] × [0, 1]2, we have Zb Zd Z1 Z1 a p 1 + b p 1 ¤ p1 1 , 2 0 0 a c £ p c p 2 + d p 2 ¤ p1 ´ 2 r u 2 + (1 − r ) 2 ³£ a p 1 + b p 1 ¤ p1 £ p 2 c p 2 + d p 2 ¤ p1 ´ 1 , ru 2 d t d r d xd u ×g t x p 1 + (1 − t ) + (1 − r ) 2 2 Zb Zd Z1 p1 p2 p 1 −1 p 2 −1 x u f (x, u)g (x, u)d xd u h 1 (t )k 1 (t )d t ≤ p (b 1 − a p 1 )(d p 2 − c p 2 ) a c 0 Z1 Zb ³ ³ £ c p2 + d p2 ¤ 1 ´ ³ £ c p2 + d p2 ¤ 1 ´ p1 p 1 −1 p2 p2 h 2 (t )k 2 (t )d t + p x f x, g x, dx b 1 − a p1 a 2 2 0 Zd ³£ a p 1 + b p 1 ¤ 1 ³£ a p 1 + b p 1 ¤ 1 ´ ³£ a p 1 + b p 1 ¤ 1 ´ p2 p 2 −1 p1 p1 p1 , u g , u d u + f , u f + p d 2 − c p2 c 2 2 2 Z1 Z £ c p 2 + d p 2 ¤ 1 ´ ³£ a p 1 + b p 1 ¤ 1 £ c p 2 + d p 2 ¤ 1 ´´ 1 p2 p1 p2 h 2 (t )k 2 (t )d t h 1 (t )k 1 (t )d t g , 2 2 2 0 0 p1p2 p p 1 (b − a 1 )(d p 2 − c p 2 ) x p 1 −1 u p 2 −1 f ³£ t x p 1 + (1 − t ) HERMITE-HADAMARD TYPE INEQUALITIES 315 Zb Zd ³ £ c p2 + d p2 ¤ 1 ´ p1 p2 p 1 −1 p 2 −1 p2 x u f x, g (x, u)d xd u (b p 1 − a p 1 )(d p 2 − c p 2 ) a c 2 Zb Zd ³ £ c p2 + d p2 ¤ 1 ´ p1p2 p 1 −1 p 2 −1 p2 d xd u x u f (x, u)g x, + p (b 1 − a p 1 )(d p 2 − c p 2 ) a c 2 Zd ³£ a p 1 + b p 1 ¤ 1 £ c p 2 + d p 2 ¤ 1 ´ ³£ a p 1 + b p 1 ¤ 1 ´ p2 p 2 −1 p1 p1 p2 , u d u f , u g + p d 2 − c p2 c 2 2 2 Zd ³£ a p 1 + b p 1 ¤ 1 ´ ³£ a p 1 + b p 1 ¤ 1 £ c p 2 + d p 2 ¤ 1 ´´ p2 p 2 −1 p1 p1 p2 u f + p , u d u g , d 2 − c p2 c 2 2 2 Z1 Z1 h 2 (t )k 2 (1 − t )d t h 1 (t )k 1 (t )d t ³ + 0 0 Zb Zd ³£ a p 1 + b p 1 ¤ 1 ´ p1 p2 p 1 −1 p 2 −1 p1 x u f (x, u)g , u d xd u (b p 1 − a p 1 )(d p 2 − c p 2 ) a c 2 Zb Zd ³£ a p 1 + b p 1 ¤ 1 ´ p1p2 p 1 −1 p 2 −1 p1 , u g (x, u)d xd u + p x u f (b 1 − a p 1 )(d p 2 − c p 2 ) a c 2 Zb ³£ a p 1 + b p 1 ¤ 1 £ c p 2 + d p 2 ¤ 1 ´ ³ £ c p2 + d p2 ¤ 1 ´ p1 p 1 −1 p2 p1 p2 + p d x f x g x, , b 1 − a p1 a 2 2 2 Zb ³ £ c p2 + d p2 ¤ 1 ´ p1 p 1 −1 p2 dx x f x, + p b 1 − a p1 a 2 Z1 ³£ a p 1 + b p 1 ¤ 1 £ c p 2 + d p 2 ¤ 1 ´´ Z1 p1 p2 h 1 (t )k 1 (1 − t )d t h 2 (t )k 2 (t )d t , g 2 2 0 0 Z Z ³ b d ³£ a p 1 + b p 1 ¤ 1 £ c p 2 + d p 2 ¤ 1 ´ p1p2 p 1 −1 p 2 −1 p1 p2 + p , x u g (x, u)d xd u f (b 1 − a p 1 )(d p 2 − c p 2 ) a c 2 2 Zb ³£ a p 1 + b p 1 ¤ 1 ´ ³ £ c p 2 + d p 2 ¤ 1 ´ Zd p2 p1 u p 2 −1 f dx ,u du x p 1 −1 g x, + 2 2 c a Zb ³£ a p 1 + b p 1 ¤ 1 ´ ³ £ c p 2 + d p 2 ¤ 1 ´ Zd p 1 −1 p2 p1 dx ,u du u p 2 −1 g + x f x, 2 2 c a Zb Zd ³£ a p 1 + b p 1 ¤ 1 £ c p 2 + d p 2 ¤ 1 ´´ p1 p2 x p 1 −1 u p 2 −1 f (x, u)d xd u g + , 2 2 a c Z1 Z1 h 1 (t )k 1 (1 − t )d t h 2 (t )k 2 (1 − t )d t . (3.44) ³ + 0 0 Applying Theorems 1.1 and 1.2 to 2nd-16th term of right hand side of the inequality (3.44), we have Zb £ c p2 + d p2 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p 1 −1 p2 p2 )g (x, )d x x f (x, b p1 − a p1 a 2 2 ³ £ c p 2 +d p 2 ¤ 1 £ c p 2 +d p 2 ¤ 1 £ c p 2 +d p 2 ¤ 1 £ c p 2 +d p 2 ¤ 1 ´ p2 p2 p2 p2 ≤ f (a, )g (a, )+ f (b, )g (b, ) 2 2 2 2 Z1 ³ £ c p2 + d p2 ¤ 1 £ c p2 + d p2 ¤ 1 p2 p2 )g (b, ) × h 1 (t )k 1 (t )d t + f (a, 2 2 0 Z £ c p2 + d p2 ¤ 1 £ c p2 + d p2 ¤ 1 ´ 1 p2 p2 h 1 (t )k 1 (1 − t )d t + f (b, )g (a, ) 2 2 0 316 WENGUI YANG ³ ´ 1 1 ≤ [ f (a, c) + f (a, d )][g (a, c) + g (a, d )] + [ f (b, c) + f (b, d )][g (b, c) + g (b, d )] 4h 2 ( )k 2 ( ) 2 2 Z1 Z1 Z1 h 1 (t )k 1(t )d t k 2 (t )d t h 2 (t )d t 0 0 0 ³ ´ + [ f (a, c) + f (a, d )][g (b, c) + g (b, d )] + [ f (a, c) + f (a, d )][g (b, c) + g (b, d )] Z1 Z1 Z1 1 1 ×4h 2 ( )k 2 ( ) h 1 (t )k 1 (1 − t )d t k 2 (t )d t h 2 (t ) 2 2 0 0 0 Z Z1 Z1 1 1 1 = (M 1 (a, b, c, d )+M 2 (a, b, c, d ))4h 2 ( )k 2 ( ) h 2 (t ) k 2 (t )d t d t h 1 (t )k 1 (t )d t 2 2 0 0 0 Z1 Z1 Z1 1 1 h 1 (t )k 1 (1 − t )d t , +(M 3 (a, b, c, d ) + M 4 (a, b, c, d ))4h 2 ( )k 2 ( ) k 2 (t )d t d t h 2 (t ) 2 2 0 0 0 (3.45) Zd £ a p1 + b p1 ¤ 1 £ a p1 + b p1 ¤ 1 p2 p1 p1 , u)( , u)d u u p 2 −1 f ( p p d 2 −c 2 c 2 2 ³ £ a p 1 +b p 1 ¤ 1 ³£ a p 1 +b p 1 ¤ 1 ´ ³£ a p 1 +b p 1 ¤ 1 ´ ³£ a p 1 +b p 1 ¤ 1 ´´ p1 p1 p1 p1 ≤ f( , c)g ,c +f ,d g ,d 2 2 2 2 Z1 ³ £ a p1 + b p1 ¤ 1 £ a p1 + b p1 ¤ 1 p1 p1 , c)g ( ,d) × h 2 (t )k 2 (t )d t + f ( 2 2 0 ´ Z1 £ a p1 + b p1 ¤ 1 £ a p1 + b p1 ¤ 1 p1 p1 +f ( h 2 (t )k 2 (1 − t )d t , d )g ( , c) 2 2 0 ³ ´ 1 1 ≤ [ f (a, c) + f (b, c)][g (a, c) + g (b, c)] + [ f (a, d ) + f (b, d )][g (a, d ) + g (b, d )] 4h 1 ( )k 1 ( ) 2 2 Z1 Z1 Z1 h 1 (t ) k 1 (t )d t h 2 (t )k 2 (t )d t 0 0 0 ³ ´ + f (a, c) + f (b, c)][g (a, d ) + g (b, d )] + [ f (a, d ) + f (b, d )][g (a, c) + g (b, c)] Z1 Z1 Z1 1 1 ×4h 1 ( )k 1 ( ) h 1 (t ) k 1 (t )d t h 2 (t )k 2 (1 − t )d t 2 2 0 0 0 Z1 Z1 Z1 1 1 h 1 (t ) k 1 (t )d t h 2 (t )k 2 (t )d t = (M 1 (a, b, c, d ) + M 3 (a, b, c, d ))4h 1( )k 1 ( ) 2 2 0 0 0 Z1 Z1 Z 1 1 1 +(M 2 (a, b, c, d )+M 4 (a, b, c, d ))4h 1 ( )k 1 ( ) h 1 (t ) k 1 (t )d t h 2 (t )k 2 (1−t )d t , (3.46) 2 2 0 0 0 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p2 p1 p2 f( , )g ( , ) 2 2 2 2 Z1 Z1 1 1 ≤ [ f (a, c) + f (a, d ) + f (b, c) + f (b, d )]4h 1 ( )h 2 ( ) h 1 (t )d t h 2 (t )d t 2 2 0 0 Z1 Z1 1 1 k 2 (t )d t k 1 (t )d t ×[g (a, c) + g (a, d ) + g (b, c) + g (b, d )]4k 1 ( )k 2 ( ) 2 2 0 0 = [M 1 (a, b, c, d ) + M 2 (a, b, c, d ) + M 3(a, b, c, d ) + M 4 (a, b, c, d )] Z1 Z1 Z1 Z1 1 1 1 1 ×16h 1 ( )h 2 ( )k 1 ( )k 2 ( ) k 2 (t )d t , k 1 (t )d t h 2 (t )d t h 1 (t )d t 2 2 2 2 0 0 0 0 (3.47) HERMITE-HADAMARD TYPE INEQUALITIES 317 Zb Zd £ c p2 + d p2 ¤ 1 p1 p2 p 1 −1 p 2 −1 p2 x u f (x, )g (x, u)d xd u (b p 1 − a p 1 )(d p 2 − c p 2 ) a c 2 Zd ³ £ c p2 + d p2 ¤ 1 p2 p 2 −1 p2 )g (a, u) f (a, u ≤ p d 2 − c p2 c 2 Z ´ 1 £ c p2 + d p2 ¤ 1 p2 h 1 (t )k 1 (t )d t )g (b, u) d u + f (b, 2 0 Zd ³ ³ £ c p2 + d p2 ¤ 1 ´ p2 p 2 −1 p2 f a, u + p g (b, u) d 2 − c p2 c 2 ´ Z1 ³ £ c p2 + d p2 ¤ 1 ´ p2 h 1 (t )k 1 (1 − t )d t g (a, u) d u + f b, 2 0 Z1 Z1 ³ 1 k 2 (t )d t h 2 (t )d t [g (a, c) + g (a, d )] ≤ [ f (a, c) + f (a, d )]2h 2( ) 2 0 0 Z1 Z1 ´ Z1 1 h 2 (t )d t [g (b, c) + g (b, d )] k 2 (t )d t d u +[ f (b, c) + f (b, d )]2h 2 ( ) h 1 (t )k 1 (t )d t 2 0 0 0 Z1 Z1 ³ 1 + [ f (a, c) + f (a, d )]2h 2 ( ) k 2 (t )d t h 2 (t )d t [g (b, c) + g (b, d )] 2 0 0 Z1 Z1 ´ Z1 1 h 1 (t )k 1 (1 − t )d t k 2 (t )d t h 2 (t )d t [g (a, c) + g (a, d )] +[ f (b, c) + f (b, d )]2h 2 ( ) 2 0 0 0 Z1 Z1 Z1 1 h 2 (t )d t k 2 (t )d t = (M 1 (a, b, c, d ) + M 2 (a, b, c, d ))2h 2 ( ) h 1 (t )k 1 (t )d t 2 0 0 0 Z1 Z1 Z1 1 h 1 (t )k 1 (1 − t )d t , (3.48) +(M 3 (a, b, c, d ) + M 4(a, b, c, d ))2h 2 ( ) k 2 (t )d t h 2 (t )d t 2 0 0 0 Zb Zd £ c p2 + d p2 ¤ 1 p1 p2 p 1 −1 p 2 −1 p2 x u f (x, u)g (x, )d xd u (b p 1 − a p 1 )(d p 2 − c p 2 ) a c 2 Zd ³ £ c p2 + d p2 ¤ 1 p2 p 2 −1 p2 ) f (a, u)g (a, u ≤ p d 2 − c p2 c 2 ³ £ c p 2 + d p 2 ¤ 1 ´´ Z1 p2 + f (b, u)g b, du h 1 (t )k 1(t )d t 2 0 Zd ³ £ c p2 + d p2 ¤ 1 p2 p 2 −1 p2 f (a, u)g (b, u ) + p d 2 − c p2 c 2 Z ³ £ c p 2 + d p 2 ¤ 1 ´´ 1 p2 + f (b, u)g a, h 1 (t )k 1 (1 − t )d t du 2 0 Z1 Z ³ 1 1 k 2 (t )d t h 2 (t )d t [g (a, c) + g (a, d )]2k 2( ) ≤ [ f (a, c) + f (a, d )] 2 0 0 Z1 Z1 ´ Z1 1 k 2 (t )d t d u h 1 (t )k 1(t )d t +[ f (b, c) + f (b, d )] h 2 (t )d t [g (b, c) + g (b, d )]2k 2( ) 2 0 0 0 Z1 Z1 ³ 1 k 2 (t )d t h 2 (t )d t [g (b, c) + g (b, d )]2k 2( ) + [ f (a, c) + f (a, d )] 2 0 0 Z Z1 ´ Z1 1 1 h 1 (t )k 1 (1 − t )d t k 2 (t )d t h 2 (t )d t [g (a, c) + g (a, d )]2k 2( ) +[ f (b, c) + f (b, d )] 2 0 0 0 318 WENGUI YANG Z1 Z1 Z1 1 h 1 (t )k 1 (t )d t k 2 (t )d t = (M 1 (a, b, c, d ) + M 2 (a, b, c, d ))2k 2( ) h 2 (t )d t 2 0 0 0 Z1 Z1 Z1 1 +(M 3 (a, b, c, d ) + M 4 (a, b, c, d ))2k 2( ) h 1 (t )k 1 (1 − t )d t , (3.49) k 2 (t )d t h 2 (t )d t 2 0 0 0 Zd ³£ a p 1 + b p 1 ¤ 1 £ c p 2 + d p 2 ¤ 1 ´ £ a p1 + b p1 ¤ 1 p2 p 2 −1 p1 p1 p2 , u)d u f , u g ( d p2 − c p2 c 2 2 2 ³ ³£ a p 1 + b p 1 ¤ 1 ´ ³£ a p 1 + b p 1 ¤ 1 £ c p 2 + d p 2 ¤ 1 ´ ´ Z1 £ a p1 + b p1 ¤ 1 p1 p1 p1 p2 ≤ g k 2 (t )d t f ,c + g( ,d) , 2 2 2 2 0 Z1 Z1 1 k 2 (t )d t k 1 (t )d t ≤ [g (a, c) + g (b, c) + g (a, d ) + g (b, d )]2k 1 ( ) 2 0 0 Z1 Z1 1 1 h 1 (t )d t ×[ f (a, c) + f (b, c) + f (a, d ) + f (b, d )]4h 1 ( )h 2 ( ) h 2 (t )d t 2 2 0 0 = [M 1 (a, b, c, d ) + M 2 (a, b, c, d ) + M 3(a, b, c, d ) + M 4 (a, b, c, d )] Z1 Z1 Z1 Z1 1 1 1 h 1 (t )d t h 2 (t )d t k 1 (t )d t k 2 (t )d t , (3.50) ×8h 1 ( )h 2 ( )k 1 ( ) 2 2 2 0 0 0 0 Zb £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 £ a p1 + b p1 ¤ 1 p2 p 1 −1 p1 p1 p2 , u)d xg ( , ) x f ( d p2 − c p2 a 2 2 2 ≤ [M 1 (a, b, c, d ) + M 2 (a, b, c, d ) + M 3(a, b, c, d ) + M 4 (a, b, c, d )] Z1 Z1 Z1 Z1 1 1 1 h 1 (t )d t h 2 (t )d t k 1 (t )d t k 2 (t )d t , (3.51) ×8h 1 ( )k 1 ( )k 2 ( ) 2 2 2 0 0 0 0 Zb Zd £ a p1 + b p1 ¤ 1 p1p2 p 1 −1 p 2 −1 p1 x u f (x, u)g ( , u)d xd u (b p 1 − a p 1 )(d p 2 − c p 2 ) a c 2 Zb ³ £ a p1 + b p1 ¤ 1 p1 p 1 −1 p1 , c) x f (x, c)g ( ≤ p b 1 − a p1 a 2 ´ Z1 £ a p1 + b p1 ¤ 1 p1 ,d) dx h 2 (t )k 2 (t )d t + f (x, d )g ( 2 0 Zb ³ £ a p1 + b p1 ¤ 1 p1 p 1 −1 p1 + p ,d) f (x, c)g ( x b 1 − a p1 a 2 Z ´ 1 £ a p1 + b p1 ¤ 1 p1 h 2 (t )k 2 (1 − t )d t , c) d x + f (x, d )g ( 2 0 Z1 Z1 ³ 1 k 1 (t )d t ≤ [ f (a, c) + f (b, c)] h 1 (t )d t [g (a, c) + g (b, c)]2k 1( ) 2 0 0 Z1 Z1 ´ Z1 1 h 1 (t )d t [g (a, d ) + g (b, d )]2k 1( ) +[ f (a, d ) + f (b, d )] h 2 (t )k 2 (t )d t k 1 (t )d t d x 2 0 0 0 Z1 Z1 ³ 1 k 1 (t )d t h 1 (t )d t [g (a, d ) + g (b, d )]2k 1( ) + [ f (a, c) + f (b, c)] 2 0 0 Z1 Z1 ´ Z1 1 k 1 (t )d t d x +[ f (a, d ) + f (b, d )] h 2 (t )k 2 (1 − t )d t h 1 (t )d t [g (a, c) + g (b, c)]2k 1( ) 2 0 0 0 Z1 Z1 Z1 1 h 2 (t )k 2 (t )d t k 1 (t )d t h 1 (t )d t = (M 1 (a, b, c, d ) + M 3 (a, b, c, d ))2k 1( ) 2 0 0 0 HERMITE-HADAMARD TYPE INEQUALITIES 319 Z1 Z1 Z 1 1 h 2 (t )k 2 (1 − t )d t , (3.52) k 1 (t )d t +(M 2 (a, b, c, d ) + M 4(a, b, c, d )2k 1 ( ) h 1 (t )d t 2 0 0 0 Zb Zd £ a p1 + b p1 ¤ 1 p1 p2 p 1 −1 p 2 −1 p1 x u f ( , u)g (x, u)d xd u (b p 1 − a p 1 )(d p 2 − c p 2 ) a c 2 Z1 Z1 Z1 1 h 1 (t )d t ≤ (M 1 (a, b, c, d ) + M 3 (a, b, c, d ))2h 1 ( ) k 1 (t )d t h 2 (t )k 2 (t )d t 2 0 0 0 Z1 Z1 Z1 1 h 2 (t )k 2(1 − t )d t , (3.53) k 1 (t )d t h 1 (t )d t +(M 2 (a, b, c, d ) + M 4(a, b, c, d )2h 1 ( ) 2 0 0 0 Zb £ c p2 + d p2 ¤ 1 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p 1 −1 p2 p1 p2 x g (x, )d x f ( , ) b p1 − a p1 a 2 2 2 Z ³ £ c p2 + d p2 ¤ 1 ´ 1 £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 £ c p2 + d p2 ¤ 1 p2 p2 p1 p2 ) + g (b, ) , ) k 1 (t )d t f ( ≤ g (a, 2 2 2 2 0 Z1 Z1 1 k 2 (t )d t k 1 (t )d t ≤ [g (a, c) + g (b, c) + g (a, d ) + g (b, d )]2k 2 ( ) 2 0 0 Z1 Z1 1 1 ×[ f (a, c) + f (b, c) + f (a, d ) + f (b, d )]4h 1 ( )h 2 ( ) h 2 (t )d t h 1 (t )d t 2 2 0 0 = [M 1 (a, b, c, d ) + M 2(a, b, c, d ) + M 3 (a, b, c, d ) + M 4 (a, b, c, d )] Z1 Z1 Z1 Z1 1 1 1 k 2 (t )d t , (3.54) k 1 (t )d t h 2 (t )d t h 1 (t )d t ×8h 1 ( )h 2 ( )k 2 ( ) 2 2 2 0 0 0 0 Zb £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p 1 −1 p2 p1 p2 )d xg ( , ) x f (x, b p1 − a p1 a 2 2 2 ≤ [M 1 (a, b, c, d ) + M 2(a, b, c, d ) + M 3 (a, b, c, d ) + M 4 (a, b, c, d )] Z1 Z1 Z1 Z1 1 1 1 h 1 (t )d t h 2 (t )d t k 1 (t )d t ×8h 2 ( )k 1 ( )k 2 ( ) k 2 (t )d t , (3.55) 2 2 2 0 0 0 0 Zd Zb £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p2 p 2 −1 p 1 −1 p2 p1 u f ( )d x , u)d u x g (x, (b p 1 − a p 1 )(d p 2 − c p 2 ) a 2 2 c ≤ [M 1 (a, b, c, d ) + M 2(a, b, c, d ) + M 3 (a, b, c, d ) + M 4 (a, b, c, d )] Z1 Z1 Z1 Z1 1 1 ×4h 1 ( )k 2 ( ) k 2 (t )d t , (3.56) k 1 (t )d t h 2 (t )d t h 1 (t )d t 2 2 0 0 0 0 Zd Zb £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p2 p 2 −1 p 1 −1 p2 p1 )d x , u)d u u g ( x f (x, (b p 1 − a p 1 )(d p 2 − c p 2 ) a 2 2 c ≤ [M 1 (a, b, c, d ) + M 2(a, b, c, d ) + M 3 (a, b, c, d ) + M 4 (a, b, c, d )] Z1 Z1 Z1 Z1 1 1 ×4h 2 ( )k 1 ( ) h 1 (t )d t h 2 (t )d t k 1 (t )d t k 2 (t )d t , (3.57) 2 2 0 0 0 0 Zb Zd £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p2 p1 p2 x p 1 −1 u p 2 −1 g (x, u)d xd u f ( , ) p p p p (b 1 − a 1 )(d 2 − c 2 ) a c 2 2 ≤ [M 1 (a, b, c, d ) + M 2(a, b, c, d ) + M 3 (a, b, c, d ) + M 4 (a, b, c, d )] Z1 Z1 Z1 Z1 1 1 k 2 (t )d t , (3.58) k 1 (t )d t h 2 (t )d t h 1 (t )d t ×4h 1 ( )h 2 ( ) 2 2 0 0 0 0 320 WENGUI YANG £ a p1 + b p1 ¤ 1 £ c p2 + d p2 ¤ 1 p1 p2 x p 1 −1 u p 2 −1 f (x, u)d xd u g ( , ) 2 2 a c Z1 1 1 h 1 (t )d t ≤ [M 1 (a, b, c, d ) + M 2 (a, b, c, d ) + M 3(a, b, c, d ) + M 4 (a, b, c, d )]4k 1( )k 2 ( ) 2 2 0 Z1 Z1 Z1 k 2 (t )d t . (3.59) k 1 (t )d t h 2 (t )d t Zb Zd p1p2 p p (b 1 − a 1 )(d p 2 − c p 2 ) 0 0 0 Using the inequalities (3.45)-(3.59) in (3.44), we get the desired result (3.40). Remark 7. Suppose that f , g : ∆ → [0, ∞) are convex functions on the co-ordinates on ∆. Then one has the inequality: 9 (b − a)(d − c) Zb Zd Z1 Z1 a c 0 0 f (t x + (1 − t ) c +d a +b , r u + (1 − r ) ) 2 2 a +b c +d ×g (t x + (1 − t ) , r u + (1 − r ) )d t d r d xd u 2 2 Zb Zd 1 37 11 29 ≤ f (x, y)g (x, y)d xd y + L(a, b, c, d ) + M (a, b, c, d ) + N (a, b, c, d ), (b − a)(d − c) a c 48 16 48 where L(a, b, c, d ), M (a, b, c, d ) and N (a, b, c, d ) are defined in Theorem 1.5. Remark 8. Suppose that f , g : ∆ → [0, ∞) are s-convex functions on the co-ordinates on ∆. Then one has the inequality: (1 + 2s)2 (b − a)(d − c) Zb Zd Z1 Z1 a c 0 0 f (t x + (1 − t ) a +b c +d , r u + (1 − r ) ) 2 2 a +b c +d ×g (t x + (1 − t ) , r u + (1 − r ) )d t d r d xd u 2 2 Zb Zd 1 ≤ f (x, y)g (x, y)d xd y + Π1L(a, b, c, d ) + Π2 M (a, b, c, d ) + Π3 N (a, b, c, d ), (b − a)(d − c) a c where L(a, b, c, d ), M (a, b, c, d ) and N (a, b, c, d ) are defined in Theorem 1.5, and 23−2s (2s [Γ(1 + s)]2 + Γ(1 + 2s)) 24−4s (21+s [Γ(1 + s)]2 + (1 + 4s )Γ(1 + 2s)) + , (1 + s)2 (1 + 2s)Γ(1 + 2s) (1 + s)4 Γ(1 + 2s) 22−2s (1 + [Γ(1 + s)]2 )(2s [Γ(1 + s)]2 + Γ(1 + 2s)) 24−4s (21+s [Γ(1 + s)]2 + (1 + 4s )Γ(1 + 2s)) + , Π2 = (1 + s)2 (1 + 2s)Γ(1 + 2s) (1 + s)4 Γ(1 + 2s) 23−2s [Γ(1 + s)]2 (2s [Γ(1 + s)]2 + Γ(1 + 2s)) 24−4s (21+s [Γ(1 + s)]2 + (1 + 4s )Γ(1 + 2s)) + . Π3 = (1 + s)2 (1 + 2s)[Γ(1 + 2s)]2 (1 + s)4 Γ(1 + 2s) Π1 = Remark 9. Suppose that f , g : ∆ → [0, ∞) are s 1-convex and s 2 -convex functions on the coordinates on ∆, respectively. Then one has the inequality: (1 + s 1 + s 2 )2 (b − a)(d − c) Zb Zd Z1 Z1 a c 0 0 f (t x + (1 − t ) a +b c +d , r u + (1 − r ) ) 2 2 a +b c +d ×g (t x + (1 − t ) , r u + (1 − r ) )d t d r d xd u 2 2 HERMITE-HADAMARD TYPE INEQUALITIES ≤ 1 (b − a)(d − c) Zb Zd a c 321 f (x, y)g (x, y)d xd y + Σ1L(a, b, c, d ) + Σ2 M (a, b, c, d ) + Σ3 N (a, b, c, d ), where L(a, b, c, d ), M (a, b, c, d ) and N (a, b, c, d ) are defined in Theorem 1.5, and 23−s1 −s2 Γ(1 + s 1 + s 2 ) + (22−s1 + 22−s2 )Γ(1 + s 1 )Γ(1 + s 1 ) (1 + s 1 )(1 + s 2 )(1 + s 1 + s 2 )Γ(1 + s 1 + s 2 ) (24−2s1 −2s2 + 22−2s1 + 23−s1 −s2 + 22−2s2 )Γ(1 + s 1 + s 2 ) + 23−s1 −s2 (2−s1 + 2−s2 )Γ(1 + s 1 )Γ(1 + s 1 ) , + (1 + s 1 )2 (1 + s 2 )2 Γ(1 + s 1 + s 2 ) ¡ ¢ (1 + Γ(1 + s 1 )Γ(1 + s 2 )) 22−s1 −s2 Γ(1 + s 1 + s 2 ) + (21−s1 + 21−s2 )Γ(1 + s 1 )Γ(1 + s 1 ) Σ2 = (1 + s 1 )(1 + s 2 )(1 + s 1 + s 2 )Γ(1 + s 1 + s 2 ) (24−2s1 −2s2 + 22−2s1 + 23−s1 −s2 + 22−2s2 )Γ(1 + s 1 + s 2 ) + 23−s1 −s2 (2−s1 + 2−s2 )Γ(1 + s 1 )Γ(1 + s 1 ) + , (1 + s 1 )2 (1 + s 2 )2 Γ(1 + s 1 + s 2 ) ¡ ¢ Γ(1 + s 1 )Γ(1 + s 2 ) 23−s1 −s2 Γ(1 + s 1 + s 2 ) + (22−s1 + 22−s2 )Γ(1 + s 1 )Γ(1 + s 1 ) Σ3 = (1 + s 1 )(1 + s 2 )(1 + s 1 + s 2 )Γ(1 + s 1 + s 2 ) 4−2s 1 −2s 2 2−2s 1 3−s 1 −s 2 (2 +2 +2 + 22−2s2 )Γ(1 + s 1 + s 2 ) + 23−s1 −s2 (2−s1 + 2−s2 )Γ(1 + s 1 )Γ(1 + s 1 ) + . (1 + s 1 )2 (1 + s 2 )2 Γ(1 + s 1 + s 2 ) Σ1 = References [1] J. Pečarić, F. Proschan and Y. L. Tong, Convex functions, partial ordering and statistical applications, Academic Press, New York, 1991. [2] S. S. Dragomir and C. E. M. Pearce, Selected Topics on Hermite-Hadamard Inequalities and Applications, RGMIA Monographs, Victoria University, 2000. [3] S. S. Dragomir, Some new Hermite-Hadamard’s type fractional integral inequalities, J. Comput. Anal. Appl., 18(2015), 655–661. [4] M. A. Noor, G. Cristescu and M. U. Awan, Generalized fractional Hermite-Hadamard inequalities for twice differentiable s-convex functions, Filomat 29(2015), 507–585. [5] I. Iscan and S. Wu, Hermite-Hadamard type inequalities for harmonically convex functions via fractional integrals, Appl. Math. Comput., 238 (2014), 237–244. [6] F. Chen and S. Wu, Some Hermite-Hadamard type inequalities for harmonically s-convex functions, Sci. World J., 2014 (2014), 279158. [7] W. Li and F. Qi, Some Hermite-Hadamard type inequalities for functions whose n-th derivatives are (α,m)convex, Filomat, 27 (2013), 1575–1582. [8] L. Chun and F. Qi, Integral inequalities of Hermite-Hadamard type for functions whose third derivatives are convex, J. Inequal. Appl., 2013 (2013), 451. [9] J. Wang, J. Deng and M. Fekan, Hermite-Hadamard-type inequalities for r -convex functions based on the use of Riemann-Liouville fractional integrals, Ukrain. Math. J., 65 (2013), 193–211. [10] J. Wang et al., Hermite-Hadamard-type inequalities for Riemann-Liouville fractional integrals via two kinds of convexity, Appl. Anal., 92 (2013), 2241–2253. [11] M. Z. Sarikaya et al., Hermite-Hadamard’s inequalities for fractional integrals and related fractional inequalities, Math. Comput. Modelling, 57 (2013), 2403–2407. [12] E. Set et al., On new inequalities of Hermite-Hadamard-Fejer type for convex functions via fractional integrals, Appl. Math. Comput., 259 (2015), 875–881. [13] S. S. Dragomir and S. Fitzpatrick, The Hadamard’s inequality for s-convex functions in the second sense, Demonstr. Math., 32(1999), 687–696. 322 WENGUI YANG [14] M. Z. Sarikaya and A. Saglam and H. Yildirim, On some Hadamard-type inequalities for h-convex functions, J. Math. Inequal., 2(2008), 335–341. [15] Z. Fang and R. Shi, On the (p,h)-convex function and some integral inequalities, J. Inequal. Appl., 2014 (2014), 45. [16] S. S. Dragomir, On the Hadamard’s inequality for convex functions on the co-ordinates in a rectangle from the plane, Taiwan. J. Math., 5 (2001), 775–788. [17] M. Alomari and M. Darus, The Hadamard’s inequality for s-convex function of 2-variables on the co-ordinates, Int. J. Math. Anal., 2 (2008), 629–638. [18] S. Bai and F. Qi, Some inequalities for (s 1 ,m 1 )-(s 2 ,m 2 )-convex functions on the co-ordinates, Glob. J. Math. Anal., 1 (2013), 22–28. [19] M. E. Ödemir, C. Yildiz and A. O. Akademir, On the co-ordinated convex functions, Appl. Math. Infor. Sci., 8 (2014), 1085-1091. [20] D. Y. Hwang, K. L. Tseng and G. S. Yang, Some Hadamard’s inequalities for co-ordinated convex functions in a rectangle from the plane, Taiwan. J. Math., 11 (2007), 63–73. [21] B. Xi, J. Hua and F. Qi, Hermite-Hadamard type inequalities for extended s-convex functions on the coordinates in a rectangle, J. Appl. Anal., 20 (2014), 29–39. [22] M. A. Latif and S. S. Dragomir, On some new inequalities for differentiable co-ordinated convex functions, J. Inequal. Appl., 2012 (2012), 28. [23] M. E. Ödemir, M. A. Latif and A. O. Akademir, On some Hadamard-type inequalities for product of two sconvex functions on the co-ordinates, J. Inequal. Appl., 2012 (2012), 21. [24] M. A. Noor, K. I. Noor and M. U. Awan, Integral inequalities for coordinated Harmonically convex functions, Complex Var. Elliptic Equ. (2014), http://dx.doi.org/10.1080/17476933.2014.976814. [25] M. Matloka, On some Hadamard-type inequalities for (h 1 ,h 2 )-preinvex functions on the co-ordinates, J. Inequal. Appl., 2013 (2013), 227. [26] M. A. Latif and M. Alomari, Hadamard-type inequalities for product two convex functions on the co-ordinates, Int. Math. Forum, 4 (2009), 2327–2338. [27] M. E. Ödemir and A. O. Akademir, On the hadamard type inequalities involving product of two convex functions on the co-ordinates, Tamkang J. Math., 46 (2015), 129–142. Ministry of Public Education, Sanmenxia Polytechnic, Sanmenxia, Henan 472000, China. E-mail: [email protected]
© Copyright 2026 Paperzz