Volume 10 (2009), Issue 2, Article 45, 7 pp. A NEW REFINEMENT OF THE HERMITE-HADAMARD INEQUALITY FOR CONVEX FUNCTIONS G. ZABANDAN zabandan@tmu.ac.ir D EPARTMENT OF M ATHEMATICS FACULTY OF M ATHEMATICAL S CIENCE AND C OMPUTER E NGINEERING T EACHER T RAINING U NIVERSITY, 599 TALEGHANI AVENUE T EHRAN 15618 IRAN. Received 13 August, 2008; accepted 12 February, 2009 Communicated by S.S. Dragomir A BSTRACT. In this paper we establish a new refinement of the Hermite-Hadamard inequality for convex functions. Key words and phrases: Hermite-Hadamard inequality. 2000 Mathematics Subject Classification. 26D15, 26D10. 1. I NTRODUCTION Let f : [a, b] → R be a convex function, then the following inequality: Z b a+b 1 f (a) + f (b) ≤ f (x)dx ≤ (1.1) f 2 b−a a 2 is known as the Hermite-Hadamard inequality [5]. In recent years there have been many extensions, generalizations and similar results of the inequality (1.1). In [2], Dragomir established the following theorem which is a refinement of the left side of (1.1). Theorem 1.1. If f : [a, b] → R is a convex function, and H is defined on [0, 1] by Z b 1 a+b H(t) = f tx + (1 − t) dx, b−a a 2 then H is convex, increasing on [0, 1], and for all t ∈ [0, 1], we have Z b a+b 1 f = H(0) ≤ H(t) ≤ H(1) = f (x)dx. 2 b−a a In [6] Yang and Hong established the following theorem which is a refinement of the right side of inequality (1.1). 226-08 2 G. Z ABANDAN Theorem 1.2. If f : [a, b] → R is a convex function, and F is defined by Z b 1+t 1−t 1+t 1−t 1 F (t) = f a+ x +f b+ x dx, 2(b − a) a 2 2 2 2 then F is convex, increasing on [0, 1], and for all t ∈ [0, 1], we have Z b 1 f (a) + f (b) f (x)dx = F (0) ≤ F (t) ≤ F (1) = . b−a a 2 In this paper we establish a refinement of the both sides of inequality (1.1). For this we first define two sequences {xn } and {yn } by n (1.2) (1.3) 2 1 X b−a b−a xn = n f a + i n − n+1 2 i=1 2 2 2n 1 X 1 b−a = n f a+ i− , 2 i=1 2 2n 2n i i i−1 i−1 1 X yn = n+1 f 1 − n a + nb + f 1− n a+ n b 2 2 2 2 2 i=1 " # n 2X −1 i i 1 f 1 − n a + nb = n+1 f (a) + f (b) + 2 2 2 2 i=1 and we prove the following a+b 1 3a + b a + 3b f = x0 ≤ f +f 2 2 4 4 = x1 ≤ · · · ≤ xn ≤ · · · Z b 1 ≤ f (x)dx ≤ · · · ≤ yn ≤ · · · ≤ y1 b−a a 1 a+b f (a) + 2f + f (b) = 4 2 f (a) + f (b) , ≤ y0 = 2 which is a new refinement of the Hermite-Hadamard inequality (1.1). For a similar discussion, see [1] or the monograph online [7, p. 19 – 22]. 2. A R EFINEMENT R ESULT In this section, using the terminologies of the introduction, we refine the Hermite-Hadamard inequality via the sequences {xn } and {yn }. Theorem 2.1. Let f be a convex function on [a, b]. Then we have Z b a+b 1 f (a) + f (b) f ≤ xn ≤ f (x)dx ≤ yn ≤ . 2 b−a a 2 J. Inequal. Pure and Appl. Math., 10(2) (2009), Art. 45, 7 pp. http://jipam.vu.edu.au/ H ERMITE -H ADAMARD I NEQUALITY 3 Proof. By the right side of Hermite-Hadamard inequality (1.1) we have Z b 1 f (x)dx b−a a b−a 2n Z 1 X a+i 2n f (x)dx = b − a i=1 a+(i−1) b−a n 2 n 2 + f a + (i − 1) b−a b−a b − a f a + i b−a 1 X 2n 2n ≤ a + i n − a − (i − 1) n 2 b − a i=1 2 2 " 2n # X 1 i i i−1 i−1 = n+1 f 1 − n a + nb + f 1− n a + n b 2 2 2 2 2 i=1 = yn . By the convexity of f we obtain 2n 1 X i i i−1 i−1 yn ≤ n+1 1 − n f (a) + n f (b) + 1 − n f (a) + n f (b) 2 2 2 2 2 i=1 " # 2n 2n X X 1 i 1 i 1 = n+1 f (a) 2 − n−1 + n + f (b) − 2 2 2 2n−1 2n i=1 i=1 1 1 2n (2n + 1) 2n 1 2n (2n + 1) 2n n+1 = n+1 f (a) 2 − n−1 + n + f (b) n−1 · − n 2 2 2 2 2 2 2 1 f (a) + f (b) , = n+1 [f (a)(2n+1 − 2n ) + f (b)(2n )] = 2 2 so Z b f (a) + f (b) 1 f (x)dx ≤ yn ≤ . b−a a 2 On the other hand, by the left side of inequality (1.1) we have b−a Z b 2n Z 2n 1 1 X a+i 2n 1 Xb−a f (x)dx = f (x)dx ≥ , b−a a b − a i=1 a+(i−1) b−a b − a i=1 2n 2n ! 2n b−a X a + i b−a + a + (i − 1) 1 b−a b−a n n 2 2 f = n f a + i n − n+1 = xn . 2 2 i=1 2 2 By the convexity of f and Jensen’s inequality we obtain 2n 1 X b−a b−a xn = n f a + i n − n+1 2 i=1 2 2 " # 2n 1 X b−a b−a a + i n − n+1 ≥f n 2 i=1 2 2 1 b − a 2n (2n + 1) b − a n n =f n 2 a+ n · − n+1 2 2 2 2 2 b−a a+b =f a+ =f . 2 2 J. Inequal. Pure and Appl. Math., 10(2) (2009), Art. 45, 7 pp. http://jipam.vu.edu.au/ 4 G. Z ABANDAN Theorem 2.2. Let f be a convex function on [a, b], then {xn } is increasing, {yn } is decreasing and Z b 1 lim xn = lim yn = f (x)dx. n→∞ n→∞ b−a a Proof. We have 2n 1 X b−a b−a xn = n f a + i n − n+1 2 i=1 2 2 n+1 2n 1 X (2 − 2i + 1)a + (2i − 1)b = n f 2 i=1 2n+1 2n 1 X 1 (2n+3 − 8i + 4)a + (8i − 4)b = n f · 2 i=1 2 2n+2 2n 1 X 1 (2n+2 + 3 − 4i)a + (4i − 3)b + (2n+2 + 1 − 4i)a + (4i − 1)b = n f · 2 i=1 2 2n+2 n+2 2n 1 X (2 + 3 − 4i)a + (4i − 3)b ≤ n+1 f 2 2n+2 i=1 n+2 2n (2 + 1 − 4i)a + (4i − 1)b 1 X f + n+1 2 2n+2 i=1 set A = {1, 3, . . . , 2n+1 − 1} and B = {2, 4, . . . , 2n+1 }, thus we obtain n+2 X n+2 2n X (2 + 3 − 4i)a + (4i − 3)b (2 + 1 − 2i)a + (2i − 1)b f = f 2n+2 2n+2 i=1 A n+2 X n+2 2n X (2 + 1 − 4i)a + (4i − 1)b (2 + 1 − 2i)a + (2i − 1)b f = f , n+2 n+2 2 2 i=1 B which implies that xn ≤ 1 " 2n+1 # X (2n+2 + 1 − 2i)a + (2i − 1)b f = xn+1 , n+2 2 A∪B so {xn } is increasing. On the other hand we have yn+1 = 1 2n+2 " f (a) + f (b) + 2 2n+1 X−1 f 1− i a+ i 2n+1 i=1 " # 2n+1 X−1 (2n+1 − i)a + ib 1 = n+2 f (a) + f (b) + 2 f . 2 2n+1 i=1 J. Inequal. Pure and Appl. Math., 10(2) (2009), Art. 45, 7 pp. 2n+1 # b http://jipam.vu.edu.au/ H ERMITE -H ADAMARD I NEQUALITY 5 Setting C = {2, 4, 6, . . . , 2n+1 − 2}, we obtain yn+1 = = = ≤ = = = # X (2n+1 − i)a + ib X (2n+1 − i)a + ib f (a) + f (b) + 2 f +2 f n+1 2n+2 2 2n+1 i∈C i∈A " n −1 n+1 2X (2 − 2i)a + 2ib 1 f (a) + f (b) + 2 f 2n+2 2n+1 i=1 n+1 # 2n X (2 − 2i + 1)a + (2i − 1)b +2 f 2n+1 i=1 " n −1 n 2X 1 (2 − i)a + ib f (a) + f (b) + 2 f 2n+2 2n i=1 # 2n n n X 1 (2 − i)a + ib + (2 − i + 1)a + (i − 1)b +2 f · 2 2n i=1 " n −1 n 2X 1 (2 − i)a + ib f (a) + f (b) + 2 f 2n+2 2n i=1 n X n # 2n 2n X (2 − i)a + ib (2 − i + 1)a + (i − 1)b + f + f n 2 2n i=1 i=1 " n −1 n −1 n 2X n 2X 1 (2 − i)a + ib (2 − i)a + ib f (a) + f (b) + 2 f + f n 2n+2 2 2n i=1 i=1 # n 2n X (2 − i + 1)a + (i − 1)b +f (b) + f (a) + f 2n i=2 " n −1 n −1 n 2X n # 2X 1 (2 − i)a + ib (2 − i)a + ib 2f (a) + 2f (b) + 3 f + f 2n+2 2n 2n i=1 i=1 " n −1 n # 2X 1 (2 − i)a + ib f (a) + f (b) + 2 f = yn , n 2n+1 2 i=1 1 " so {yn } is decreasing. For the proof of the last assertions, since f is continuous on [a, b], we use the following well known equality: Z b n b−a b−aX lim f a+i = f (x)dx. n→∞ n i=1 n a So we obtain 1 lim xn = lim yn = n→∞ n→∞ b−a Z b f (x)dx. a J. Inequal. Pure and Appl. Math., 10(2) (2009), Art. 45, 7 pp. http://jipam.vu.edu.au/ 6 G. Z ABANDAN Remark 1. Let f be a convex function on [a, b]. In conclusion, we can state that a+b 3a + b a + 3b 1 f = x0 ≤ f +f 2 2 4 4 = x1 ≤ · · · ≤ xn ≤ · · · Z b 1 ≤ f (x)dx ≤ · · · ≤ yn ≤ · · · ≤ y1 b−a a a+b 1 = f (a) + 2f + f (b) 4 2 f (a) + f (b) ≤ y0 = . 2 3. A PPLICATIONS FOR S PECIAL M EANS Recall the following means a) The arithmetic mean A(a, b) = b) The geometric neam a+b 2 (a, b > 0); √ ab G(a, b) = (a, b > 0); c) The harmonic mean H(a, b) = 1 a 2 + 1 b (a, b > 0); d) The logarithmic mean ( L(a, b) = b−a ln b−ln a a b 6= a; b = 0; (a, b > 0). We define the two new means by the following: e) The n−harmonic mean " n −1 2X 1 n+1 1 Hn (a, b) = 2 +2 i a 1 − 2n a + i=1 (n = 0, 1, 2, . . . , i b 2n + 1 b #−1 a, b > 0) f) The n−arithmetic mean " An (a, b) = 2n #−1 n 2 X i=1 1− i 2n 1 1 + 2n+1 a + i 2n − 1 2n+1 b (n = 0, 1, 2, . . . ; a, b > 0). It is clear that H0 (a, b) = H(a, b) and A0 (a, b) = A(a, b). By the above terminology we have the following simple proposition: Proposition 3.1. Let 0 < a < b < ∞. Then we have H(a, b) ≤ Hn (a, b) ≤ L(a, b) ≤ An (a, b) ≤ A(a, b), lim Hn (a, b) = lim An (a, b) = L(a, b). n→∞ Proof. Let f : [a, b] → (0, ∞), f (x) = n→∞ 1 x J. Inequal. Pure and Appl. Math., 10(2) (2009), Art. 45, 7 pp. and use Remark 1. We omit the details. http://jipam.vu.edu.au/ H ERMITE -H ADAMARD I NEQUALITY 7 R EFERENCES [1] S.S. DRAGOMIR, Some remarks on Hadamard’s inequalities for convex functions, Extracta Math., 9(2) (1994), 88–94. [2] S.S. DRAGOMIR, Two mappings in connection to Hadamard’s inequalities, J. Math. Anal. Appl., 167 (1992), 42–56. [3] S.S. DRAGOMIR AND A. McANDREW, Refinement of the Hermite-Hadamard inequality for convex functions, J. Inequal. Pure and Appl. Math., 6(5) (2005), Art. 140. [ONLINE: http: //jipam.vu.edu.au/article.php?sid=614]. [4] S.S. DRAGOMIR, Y.J. CHO AND S.S. KIM. Inequalities of Hadamard’s type for Lipschitzian mappings and their applications, J. Math. Anal. Appl., 245 (2000), 489–501. [5] J. HADAMARD, Étude sur les proprietes des fonctions entieres en particulier d’une fonction considĚerée par Riemann, J. Math. Pures Appl., 58 (1893), 171–215. [6] G.S. YANG AND M.C. HONG, A note on Hadamard’s inequality, Tamkang. J. Math., 28(1) (1997), 33–37. [7] S.S. DRAGOMIR AND C.E.M. PEARCE, Selected Topics on the Hermite Hadamard Inequality and Applications, RGMIA Monographs, Victoria University, 2000. [ONLINE: http://www.staff. vu.edu.au/RGMIA/monographs/hermite_hadamard.html]. J. Inequal. Pure and Appl. Math., 10(2) (2009), Art. 45, 7 pp. http://jipam.vu.edu.au/