Journal of Inequalities in Pure and Applied Mathematics SOME EXTENSIONS OF HILBERT’S TYPE INEQUALITY AND ITS APPLICATIONS volume 7, issue 2, article 61, 2006. YONGJIN LI AND BING HE Department of Mathematics Sun Yat-sen University Guangzhou, 510275 P. R. China. Received 28 November, 2005; accepted 19 January, 2006. Communicated by: B. Yang EMail: stslyj@mail.sysu.edu.cn EMail: hzs314@163.com Abstract Contents JJ J II I Home Page Go Back Close c 2000 Victoria University ISSN (electronic): 1443-5756 342-05 Quit Abstract By introducing parameters λ and µ, we give a generalization of the Hilbert’s type integral inequality. As applications, we give its equivalent form. 2000 Mathematics Subject Classification: 26D15. Key words: Hilbert’s integral inequality, Weight function. Some Extensions of Hilbert’s Type Inequality and its Applications The authors are grateful to the referee for the careful reading of the manuscript and for several useful remarks. The work was partially supported by the Foundation of Sun Yat-sen University Advanced Research Centre. Yongjin Li and Bing He Contents 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Main Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References Title Page 3 6 Contents JJ J II I Go Back Close Quit Page 2 of 15 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au 1. Introduction 1 p If f, g ≥ 0, p > 1, ∞, then ∞ Z ∞ Z 1 q = 1, 0 < 0 π < sin(π/p) ∞ Z Z ∞ (1.2) 0 R∞ f p (x)dx < ∞ and 0 < 0 R∞ 0 g q (x)dx < f (x)g(y) dxdy x+y (1.1) 0 + 0 f (x) dx x+y p p1 Z ∞ Z p 1q ∞ q f (x)dx g (x)dx 0 , Some Extensions of Hilbert’s Type Inequality and its Applications 0 π dy < sin(π/p) p Z ∞ Yongjin Li and Bing He f p (x)dx, 0 π where the constant factor sin(π/p) is the best possible. Inequality (1.1) is known as Hardy-Hilbert’s integral inequality (see [1]); it is important in analysis and its applications (see [4]). Under the same condition of (1.1), we have the HardyHilbert type inequality similar to (1.1): Z ∞ ∞ Z (1.3) 0 0 f (x)g(y) dxdy < pq max{x, y} Z ∞ f p (x)dx p1 Z 0 ∞ g q (x)dx 0 1q , Title Page Contents JJ J II I Go Back Close Quit Z ∞ Z (1.4) 0 0 ∞ f (x) dx max{x, y} p dy < (pq)p Z ∞ f p (x)dx, Page 3 of 15 0 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au where the constant factor pq is the best possible. The corresponding inequality for series is: ! p1 ∞ ! 1q ∞ X ∞ ∞ X X X am b n (1.5) < pq apn bqn , max{m, n} n=1 m=1 n=1 n=1 where the constant factor pq is also the best possible. In particular, when p = q = 2, we have the well-known Hilbert type inequality: 12 Z ∞ 12 Z ∞ Z ∞Z ∞ f (x)g(y) (1.6) g 2 (x)dx . dxdy < 4 f 2 (x)dx max{x, y} 0 0 0 0 In recent years, Kuang (see [3]) gave a strengthened form as: (1.7) ∞ X ∞ X am b n max{m, n} n=1 m=1 (∞ ) p1 ( ∞ ) 1q X X [pq − G(q, n)]bqn , < [pq − G(p, n)]apn n=1 n=1 where G(r, n) = r+1/3r−4/3 > 0 (r = p, q). (2n+1)1/r Yang (see [5, 8]) gave: for λ > 2 − min{p, q} Z ∞Z ∞ f (x)g(y) pqλ dxdy < (1.8) λ λ max{x , y } (p + λ − 2)(q + λ − 2) 0 0 Z ∞ p1 Z ∞ 1q (p−1)(2−λ)−1 p (q−1)(2−λ)−1 q × x f (x)dx x g (x)dx 0 0 Some Extensions of Hilbert’s Type Inequality and its Applications Yongjin Li and Bing He Title Page Contents JJ J II I Go Back Close Quit Page 4 of 15 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au and ∞ Z Z (1.9) 0 0 ∞ f (x)g(y) pqλ dxdy < max{xλ , y λ } (p + λ − 2)(q + λ − 2) Z ∞ p1 Z ∞ 1q 1−λ p 1−λ q x f (x)dx × x g (x)dx . 0 0 At the same time, Yang (see [6, 7]) considered the refinement of other types of Hilbert’s inequalities. In this paper, we give a generalization of Hilbert’s type inequality and an improvement as: Some Extensions of Hilbert’s Type Inequality and its Applications Yongjin Li and Bing He Z 0 ∞ Z 0 ∞ f (xλ )g(y µ ) dxdy max{x, y} 1q Z ∞ p1 Z ∞ 1 1 pq −1 q −1 p µ λ < 1 1 x f (x)dx x g (x)dx , 0 0 λp µq where λ > 0 and µ > 0. Title Page Contents JJ J II I Go Back Close Quit Page 5 of 15 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au 2. Main Results Lemma 2.1. Suppose r > 1, 1s + x (2.1) = 1, λ, µ, ε > 0. Then 1 ∞ Z 1 r µ −ε( 1s + λr )−1 x− λ Z 1 −1−µε 1 t r dtdx = O(1) max{1, t} 0 Proof. There exist n ∈ N which is large enough, such that 1 + 1 ε ∈ (0, µn ] and x ≥ 1, we have 1 Z 0 x− λ (ε → o+ ). −1−µε r 1 −1−µε 1 t r dt = max{1, t} Z x− λ t 1+ −1−µε r x 1 −λ dt = 0 Since for a ≥ 1 the function g(y) = 1 −1−µε r 1+ −1−µε r 1 yay ≤ 1 1+ −1−µε r x 1 −λ 1 1+ −1−1/n r x 1 −λ 1+ −1−1/n r 1 ∞ 1 µ x−ε( s + λr )−1 0< 1 Z < 0 ∞ x 1 −1 1 1 = λ x− λ Z 1+ 1 1+ −1−1/n r !2 −1−1/n r 1+ −1−µε r . Some Extensions of Hilbert’s Type Inequality and its Applications Yongjin Li and Bing He (y ∈ (0, ∞)) is decreasing, we find so Z > 0 for −1−µ 1 t r dtdx max{1, t} 1 1+ −1−1/n r x− λ . Hence relation (2.1) is valid. The lemma is proved. , Title Page Contents JJ J II I Go Back Close Quit Page 6 of 15 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au Now we study the following inequality: Theorem 2.2. Suppose f (x), g(x) ≥ 0, p > 1, p1 + 1q = 1, λ > 0, µ > 0 and Z ∞ Z ∞ 1 1 −1 p 0< x λ f (x)dx < ∞, 0 < x µ −1 g q (x)dx < ∞. 0 0 ∞ f (xλ )g(y µ ) dxdy max{x, y} 1q p1 Z ∞ Z ∞ 1 1 pq −1 −1 p q µ < 1 1 x λ f (x)dx x g (x)dx , 0 0 λp µq Then Z ∞ Z (2.2) 0 0 where the constant factor pq 1 1 λp µq is the best possible for λ = µ. Proof. By Hölder’s inequality, we have Z ∞Z ∞ f (xλ )g(y µ ) dxdy max{x, y} 0 0 ih i h Z ∞ Z ∞ x λ1 −1 f (x) y µ1 −1 g(y) 11 n 1 1o = dxdy λµ 0 0 max x λ , y µ ! p1 Z ∞Z ∞ 1 1 −1 (1− λ )p/q λ 11 x f (x) x = 1 n 1 1 o p1 λµ 0 0 y 1− µ max x λ , y µ Some Extensions of Hilbert’s Type Inequality and its Applications Yongjin Li and Bing He Title Page Contents JJ J II I Go Back Close Quit x y 1 λ 1 µ ! pq1 Page 7 of 15 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au × y 1 −1 µ g(y) n 1 1 o 1q max x λ , y µ y x ! p1 1 (1− µ ) 1 (1− λ )p/q y x Z Z 1 1 1 ∞ ∞ p( 1 −1) f p (x) x(1− λ )p/q λ n 1 1o x ≤ 1 λµ 0 0 y 1− µ max x λ , y µ Z × ∞ 0 Z ∞ 1 y q( µ −1) 0 1 µ ! pq1 dxdy 1 λ x y 1 λ g (y) y n 1 1o 1 x1− λ max x λ , y µ p1 ! 1q dxdy 1 µ 1 (1− µ )q/p q y 1 µ 1 xλ ! p1 1q dxdy . Some Extensions of Hilbert’s Type Inequality and its Applications Yongjin Li and Bing He Define the weight function ϕ(x), ψ(y) as Z ∞ n 1 1o · max x λ , y µ ϕ(x) := 0 Z ψ(y) := 0 1 ∞ 1 n 1 1o · max x λ , y µ 1 x(1− λ )p/q y 1 1− µ 1 y (1− µ )q/p 1 x1− λ 1 xλ y ! 1q dy, 1 µ 1 yµ 1 xλ x ∈ (0, ∞), Title Page Contents ! p1 dx, y ∈ (0, ∞), JJ J II I Go Back then above inequality yields Z ∞Z ∞ f (xλ )g(y µ ) dxdy max{x, y} 0 0 Z ∞ p1 Z ∞ 1q 1 1 11 −1) q q( µ p( λ −1) p ≤ ϕ(x)x f (x)dx g (y)dy . ψ(y)y λµ 0 0 Close Quit Page 8 of 15 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au 1 1 For fixed x, let y µ = x λ t, we have ϕ(x) := µx 1 (p−1)(1− λ ) ∞ Z 0 1 1 t p −1 dt max{1, t} 1 (p−1)(1− λ ) = µpqx 1 = µpqx(p−1)(1− λ ) . 1 By the same token, ψ(y) = λpqy (q−1)(1− µ ) , thus Z ∞Z ∞ f (xλ )g(y µ ) (2.3) dxdy max{x, y} 0 0 1q Z ∞ p1 Z ∞ 1 1 pq −1 −1 p q ≤ 1 1 x µ g (x)dx . x λ f (x)dx 0 0 λp µq If (2.3) takes the form of the equality, then there exist constants c and d, such that Kuang (see [2]) ! 1q ! p1 1 1 1 1 1 1 xp( λ −1) f p (x) x(1− λ )p/q x λ y q( µ −1) g q (y) y (1− µ )q/p y µ n 1 1o · c · =d 1 1 1 1 1 1 xλ x1− λ max{x λ , y µ } y 1− µ yµ max x λ , y µ a.e. on (0, ∞) × (0, ∞). Some Extensions of Hilbert’s Type Inequality and its Applications Yongjin Li and Bing He Title Page Contents JJ J II I Go Back Close Quit Then we have 1 1 cx λ f p (x) = dy µ g q (y) a.e. on (0, ∞) × (0, ∞). Page 9 of 15 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au Hence we have 1 1 cx λ f p (x) = dy µ g q (y) = constant a.e. on (0, ∞) × (0, ∞), which contradicts the facts that Z ∞ 1 0< x λ f p (x)dx < ∞ ∞ Z 0 1 y µ g q (x)dx < ∞. and 0 < 0 Some Extensions of Hilbert’s Type Inequality and its Applications Hence (2.3) takes the form of strict inequality. So we have (2.2). For 0 < ε < 12 , setting fε (x) = x −ε−1/λ p , for x ∈ [1, ∞); fε (x) = 0, for −ε−1/µ q x ∈ (0, 1), and gε (y) = y , for y ∈ [1, ∞); gε (y) = 0, for y ∈ (0, 1). Assume that the constant factor 1pq 1 in (2.2) is not the best possible, then there λp µq exists a positive number K with K < pq , such that (2.2) is valid by changing to K. We have 1 1 λp µq Z pq 1 1 λp µq ∞ 0 Z 0 ∞ f (xλ )g(y µ ) dxdy max{x, y} p1 Z Z ∞ 1 −1 p <K x λ f (x)dx 0 Since Z 0 ∞ Yongjin Li and Bing He Title Page Contents JJ J II I Go Back ∞ 1 x µ −1 g q (x)dx 0 1q = K . ε Close Quit Page 10 of 15 −1−εµ 1 t q dt = pq + o(1) max{1, t} + (ε → 0 ), J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au 1 1 setting y µ = x λ t, by (2.1), we find Z ∞Z ∞ f (xλ )g(y µ ) dxdy max{x, y} 0 0 h ih i Z ∞ Z ∞ x λ1 −1 f (x) y µ1 −1 g(y) 11 n 1 1o dxdy = λµ 0 µ 0 λ max x , y = 11 λµ Z ∞ 1 Z Z ∞ 1 ∞ Z ∞ x 1 (λ −1)+ −ε− 1 λ p 1 (µ −1)+ y n 1 1o max x λ , y µ 1 (λ −1)+ −ε− 1 λ p µ µ λ 1 −ε− µ Some Extensions of Hilbert’s Type Inequality and its Applications q dxdy 1 (µ −1)+ Yongjin Li and Bing He 1 −ε− µ 11 x (t x ) n 1 o x µtµ−1 dxdt 1 1 − λµ 1 x λ max x λ , tx λ Z Z ∞ µ −1−εµ 1 1 ∞ −ε( p1 + λq )−1 q x t dtdx = 1 λ 1 x− λ max{1, t} Z ∞ Z µ −1−εµ 1 ∞ −ε( p1 + λq 1 )−1 = x t q dtdx λ 1 max{1, t} 0 1 Z x− λ −1−εµ 1 t q dtdx − max{1, t} 0 " # 1 pq = µ + o(1) . λε p1 + λq = q µ λ Title Page Contents JJ J II I Go Back Close Quit Page 11 of 15 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au Since for ε > 0 small enough, we have " # pq 1 µ + o(1) < K. λ p1 + λq 1 1 It is obvious that λ p µ q ≤ λ p + µ q (i.e. 1 λ 1 µ 1 + λq p ≤ 1 1 1 λp µq ) by Young’s inequality. Consider the case of taking the form of the equality for Young’s inequality, we p−1 1 get µ q = λ p , i.e. λ = µ, Then " # pq pq 1 1 1 + o(1) < K. µ + o(1) = 1 λ p + λq λp µq Thus we get factor pq 1 1 λp µq pq 1 1 λp µq ≤ K, which contradicts the hypothesis. Hence the constant in (2.2) is best possible for λ = µ. Remark 1. For λ = µ, inequality (2.2) becomes Z ∞ Z (2.4) 0 0 ∞ f (xλ )g(y λ ) dxdy max{x, y} 1q Z ∞ p1 Z ∞ 1 1 pq −1 q −1 p < x λ f (x)dx x λ g (x)dx , λ 0 0 where the constant factor pq λ is the best possible. Some Extensions of Hilbert’s Type Inequality and its Applications Yongjin Li and Bing He Title Page Contents JJ J II I Go Back Close Quit Page 12 of 15 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au R∞ 1 Theorem 2.3. Suppose f ≥ 0, p > 1, λ > 0 and 0 < 0 x λ −1 f p (x)dx < ∞. Then p Z ∞ Z ∞ Z ∞ 1 f (xλ ) 1 p (2.5) dx dy < (pq) x λ −1 f (x)p dx, max{x, y} λ 0 0 0 where the constant factor λ1 (pq)p is the best possible. Inequality (2.5) is equivalent to (2.4). Proof. Setting g(y) as p−1 Z ∞ 1 −1 x λ f (x) n 1 1 o dx > 0, 0 max x λ , y λ Some Extensions of Hilbert’s Type Inequality and its Applications y ∈ (0, ∞). then by (2.4), we find p Z ∞ Z ∞ Z ∞ 1 f (xλ ) −2 −1 q p−1 dx dy λ y λ g (y)dy = λ max{x, y} 0 0 0 Z ∞Z ∞ f (xλ )g(y λ ) = dxdy max{x, y} 0 0 Z ∞ p1 Z ∞ 1q 1 1 pq ≤ x λ −1 f p (x)dx y λ −1 g q (y)dy . (2.6) λ 0 0 Hence we obtain Z ∞ Z 1 −1 q p p λ g (y)dy ≤ λ (pq) (2.7) 0< y 0 0 Yongjin Li and Bing He Title Page Contents JJ J II I Go Back Close Quit Page 13 of 15 ∞ 1 x λ −1 f p (x)dx < ∞. J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au By (2.4), both (2.6) and (2.7) take the form of strict inequality, so we have (2.5). On the other hand, suppose that (2.5) is valid. By Hölder’s inequality, we find Z ∞Z ∞ f (xλ )g(y λ ) dxdy max{x, y} 0 0 Z ∞ Z ∞ f (xλ ) = dx g(y λ )dy max{x, y} 0 0 1q p p1 Z ∞ Z ∞ Z ∞ f (xλ ) q λ (2.8) g (y )dy . dx dy ≤ max{x, y} 0 0 0 Some Extensions of Hilbert’s Type Inequality and its Applications Yongjin Li and Bing He Then by (2.5), we have (2.4). Thus (2.4) and (2.5) are equivalent. Title Page 1 (pq)p λ If the constant in (2.5) is not the best possible, by (2.8), we may get a contradiction that the constant factor in (2.4) is not the best possible. Thus we complete the proof of the theorem. Remark 2. (i) For λ = µ = 1, (2.2) and (2.5) reduce respectively to (1.3) and (1.4). It follows that (2.2) is a new extension of (1.6) and (1.3) with some parameters and the equivalent form (2.4) is a new extension of (1.4). (ii) It is amazing that (2.4) and (1.9) are different, although both of them are the extensions of (1.6) with (p, q)−parameter and the best constant factor. Contents JJ J II I Go Back Close Quit Page 14 of 15 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au References [1] G.H. HARDY, J.E. LITTLEWOOD bridge Univ Press Cambridge, 1952. AND G. POLYA, Inequalities, Cam- [2] J. KUANG, Applied Inequalities (Chinese), Hunan Jiaoyu Chubanshe, Changsha, 1993. [3] J. KUANG AND L. DEBNATH, On new generalizations of Hilbert’s inequality and their applications, J. Math. Anal. Appl., 245(1) (2000), 248– 265. [4] D.S. MITRINOVIĆ, J.E. PEČARIĆ AND A.M. FINK, Inequalities Involving Functions and their Integrals and Derivatives, Mathematics and its Applications, 53. Kluwer Academic Publishers Group, Dordrecht, 1991. [5] B. YANG, Generalization of a Hilbert type inequality and its applications (Chinese), Gongcheng Shuxue Xuebao, 21(5) (2004), 821–824. [6] B. YANG, On a Hilbert’s type inequality and its applications, Math. Appl. (Wuhan), 15(1) (2002), 52–56. Some Extensions of Hilbert’s Type Inequality and its Applications Yongjin Li and Bing He Title Page Contents JJ J II I [7] B. YANG, A new Hardy-Hilbert type inequality and its applications (Chinese), Acta Math. Sinica, 46(6) (2003), 1079–1086. Go Back [8] B. YANG, Best generalization of a Hilbert-type inequality (Chinese), J. Jilin Univ. Sci., 42(1) (2004), 30–34. Quit Close Page 15 of 15 J. Ineq. Pure and Appl. Math. 7(2) Art. 61, 2006 http://jipam.vu.edu.au