Journal of Inequalities in Pure and Applied Mathematics http://jipam.vu.edu.au/ Volume 3, Issue 4, Article 56, 2002 ON GENERALIZATIONS OF L. YANG’S INEQUALITY CHANG-JIAN ZHAO AND LOKENATH DEBNATH D EPARTMENT OF M ATHEMATICS S HANGHAI U NIVERSITY S HANGHAI 200436 T HE P EOPLE ’ S R EPUBLIC OF C HINA ; AND D EPARTMENT OF M ATHEMATICS B INZHOU T EACHERS C OLLEGE B INZHOU , S HANDONG 256604 T HE P EOPLE ’ S R EPUBLIC OF C HINA . chjzhao@163.com D EPARTMENT OF M ATHEMATICS U NIVERSITY OF T EXAS -PAN A MERICAN E DINBURG , T EXAS 78539-2999, U.S.A Debnathl@panam.edu Received 03 December, 2001; accepted 02 June, 2002 Communicated by Feng Qi A BSTRACT. A geometric inequality due to L. Yang involving cosine and sine functions is generalized. Key words and phrases: L. Yang’s inequality, Theory of distribution of functional values, Jensen inequality. 2000 Mathematics Subject Classification. 26D15. 1. I NTRODUCTION The well-known inequality due to L. Yang [1, pp. 116–118] can be stated as follows: If A > 0, B > 0, A + B ≤ π and 0 ≤ λ ≤ 1, then (1.1) cos2 λA + cos2 λB − 2 cos λA · cos λB · cos λπ ≥ sin2 λπ. The equality in (1.1) holds if and only if λ = 0 or A + B = π. L. Yang’s inequality plays an important role in the theory of distribution of values of functions. See [1]. In this paper we will generalize inequality (1.1). ISSN (electronic): 1443-5756 c 2002 Victoria University. All rights reserved. 085-01 2 C HANG - JIAN Z HAO AND L OKENATH D EBNATH 2. M AIN R ESULTS In this paper, we use the following notations and abbreviations: n n! = ; r r!(n − r)! λ λ [2] [1] xij = (π + Ai + Aj ), xij = (π − Ai − Aj ); 2 2 λ λ [4] [3] xij = (π + Ai − Aj ), xij = (π − Ai + Aj ); 2 2 2 2 Hij = cos λAi + cos λAj − 2 cos λAi · cos λAj · cos λπ. Lemma 2.1 ([2]). If Ai > 0, Aj > 0, Ai + Aj ≤ π for 1 ≤ i, j ≤ n, and −1 ≤ λ ≤ 1, then 2 Hij − sin λπ = 4 4 Y [k] sin xij . k=1 Proof. By using the following two identities cos λ (Ai + Aj ) cos λ (Ai − Aj ) = cos2 λAi + cos2 λAj − 1, cos λ (Ai + Aj ) + cos λ (Ai − Aj ) = 2 cos λAi cos λAj , it is easy to observe that Hij − sin2 λπ = cos2 λAi + cos2 λAj − 1 − 2 cos λAi cos λAj cos λπ + cos2 λπ = cos2 λπ + cos λ (Ai + Aj ) cos λ (Ai − Aj ) − [cos λ (Ai + Aj ) + cos λ (Ai − Aj )] cos λπ = [cos λπ − cos λ (Ai + Aj )] [cos λπ − cos λ (Ai − Aj )] λ λ = 4 sin (π + Ai + Aj ) sin (π − Ai − Aj ) 2 2 λ λ × sin (π + Ai − Aj ) sin (π − Ai + Aj ) . 2 2 The proof is complete. Pn Theorem 2.2. If Ai > 0, λk > 0 (i = 1, 2, . . . , n; k = 1, 2, 3, 4), i=1 Ai ≤ π, n ≥ 2 being a natural number, and −1 ≤ λ ≤ 1, then " n #2 n X X n 2 2 sin λπ ≤ (n − 1 + cos λπ) cos λAk − cos λπ cos λAi 2 i=1 k=1 (2.1) 4 P4 X λ n k k=1 ≤ sin2 λπ + sin4 θij , Q4 2 64 k=1 λk 1≤i<j≤n where [k] k=1 λk xij . P4 λ k k=1 P4 (2.2) θij = The equalities in (2.1) hold if and only if λ = 0. J. Inequal. Pure and Appl. Math., 3(4) Art. 56, 2002 http://jipam.vu.edu.au/ L. YANG ’ S I NEQUALITY 3 Proof. Since y = sin x is a continuous and convex (or concave, resp.) function on [−π, 0] (or [k] [0, π], resp.), and xij ∈ [−π, 0] (or [0, π], resp.) for −1 ≤ λ ≤ 0 (or 0 ≤ λ ≤ 1, resp.), and using Jensen’s inequality (see [3]), we observe that P4 [k] k=1 λk sin xij sin θij ≤ (or ≥ , resp.) . P4 λ k k=1 Consequently 4 X (2.3) !4 4 X (sin θij )4 ≥ λk k=1 [k] On the other hand, since λk sin −xij 1 4 (2.4) [k] λk sin xij . k=1 4 X !4 [k] ≥ 0 or λk sin xij ≥ 0, resp. , then [k] λk sin −xij v u 4 uY [k] 4 ≥t λk sin xij . k=1 k=1 v ! u 4 4 X Y u 1 [k] [k] 4 λk sin xij ≥ t or λk sin xij , resp. . 4 k=1 k=1 From (2.3) and (2.4), we obtain 4 Y (2.5) [k] sin xij 4 k=1 λk Q 44 4k=1 λk P4 ≤ k=1 sin4 θij . By using Lemma 2.1, we have 4 k=1 λk Q 64 4k=1 λk P4 2 sin λπ ≤ Hij ≤ (2.6) sin4 θij + sin2 λπ. Summing both sides of (2.6) for 1 ≤ i < j ≤ n yields X (2.7) sin2 λπ ≤ 1≤i<j≤n X Hij ≤ 1≤i<j≤n X 1≤i<j≤n ! 4 λ k=1 k sin4 θij + sin2 λπ . Q4 64 k=1 λk P4 It is not difficult to see that n sin λπ = sin2 λπ. 2 1≤i<j≤n X (2.8) 2 Direct computing yields (2.9) X 1≤i<j≤n 4 k=1 λk Q 64 4k=1 λk ! P4 sin4 θij + sin2 λπ 4 X k=1 λk sin4 θij Q4 64 k=1 λk 1≤i<j≤n P4 = J. Inequal. Pure and Appl. Math., 3(4) Art. 56, 2002 n + sin2 λπ, 2 http://jipam.vu.edu.au/ 4 C HANG - JIAN Z HAO AND L OKENATH D EBNATH and (2.10) X Hij = 1≤i<j≤n X (cos2 λAi + cos2 λAj − 2 cos λAi cos λAj cos λπ) 1≤i<j≤n = X (cos2 λAi + cos2 λAj ) − 1≤i<j≤n = (n − 1) X 2 cos λAi cos λAj cos λπ 1≤i<j≤n n X cos2 λAk − cos λπ !2 cos λAi i=1 k=1 = (n − 1 + cos λπ) n X n X cos2 λAk − cos λπ k=1 − n X cos2 λAi i=1 n X !2 cos λAi . i=1 Putting (2.8), (2.9) and (2.10) into (2.7), inequality (2.7) reduces to inequality (2.1). The proof is complete. (2.11) 1 4 (i = 1, 2, 3, 4) in the right-hand of (2.7), we find that ! 4 P4 X λ k=1 k sin4 θij + sin2 λπ Q4 64 k=1 λk 1≤i<j≤n ! 4 P4 P4 [k] 4 X k=1 λk xij k=1 λk Q = sin P + sin2 λπ 4 4 64 λ λ k=1 k k=1 k 1≤i<j≤n !!4 4 X X [k] 4 sin 1 = xij + sin2 λπ 4 1≤i<j≤n k=1 X 4 λ 2 = 4 sin π + sin λπ 2 1≤i<j≤n n λπ =4 sin2 . 2 2 Remark 2.3. If taking λi = Therefore, inequality (2.1) reduces to the following inequality (2.12) !2 n n X X n sin2 λπ ≤ (n − 1 + cos λπ) cos2 λAk − cos λπ cos λAi 2 i=1 k=1 n λ ≤4 sin2 π. 2 2 The equalities in (2.12) hold if and only if λ = 0. Letting n = 2 in (2.12), we have the following result: If A > 0, B > 0, A + B ≤ π, and −1 ≤ λ ≤ 1, then (2.13) λ sin2 λπ ≤ cos2 λA + cos2 λB − 2 cos λA cos λB cos λπ ≤ 4 sin2 π. 2 The equalities in (2.13) holds if and only if λ = 0. It is obvious that inequality (2.13) is the same as L. Yang’s inequality (1.1). J. Inequal. Pure and Appl. Math., 3(4) Art. 56, 2002 http://jipam.vu.edu.au/ L. YANG ’ S I NEQUALITY 5 P Theorem 2.4. If Ai > 0 and λk > 0 for i = 1, 2, . . . , n and k = 1, 2, 3, 4, ni=1 Ai ≤ π, n ≥ 2, −1 ≤ λ ≤ 1, then n X X n 2 0 ≤ (n − 1) cos λAk − (2.14) sin2 λπ − cos λπ cos λAi cos λAj 2 1≤i,j≤n k=1 i6=j P4 ≤ 4 k=1 λk sin4 θij , Q4 128 k=1 λk 1≤i,j≤n i6=j X where [k] k=1 λk xij . P4 λ k k=1 P4 θij = The equalities in (2.14) hold if and only if λ = 0. Proof. It follows from inequality (2.6) that 4 k=1 λk Q 64 4k=1 λk P4 2 0 ≤ Hij − sin λπ ≤ (2.15) sin4 θij . Summing both sides of (2.15) for i 6= j, first over j from 1 to n and then over i from 1 to n of the resulting inequality, then 4 P4 X X λ n k k=1 (2.16) 0≤ Hij − 2 sin2 λπ ≤ sin4 θij . Q4 2 64 λ k=1 k 1≤i,j≤n 1≤i,j≤n i6=j i6=j On the other hand n X X X (2.17) Hij = 2(n − 1) cos2 λAk − 2 cos λπ cos λAi cos λAj . 1≤i,j≤n i6=j k=1 1≤i,j≤n i6=j Putting (2.17) into (2.16), we get inequality (2.14). (2.18) 1 4 for k = 1, 2, 3, 4, then n X X n 2 0 ≤ (n − 1) cos λAk − sin2 λπ − cos λπ cos λAi cos λAj 2 1≤i,j≤n k=1 Remark 2.5. In (2.14), if λk = i6=j n λ ≤ sin4 π. 2 2 The equalities in (2.18) hold if and only if λ = 0. Letting n = 2, then (2.18) reduces to inequality (2.13). R EFERENCES [1] L. YANG, Distribution of Values and New Research, Science Press, Beijing, 1982. (Chinese). [2] Ch.-J. ZHAO, Further research of L. Yang’s inequality, J. Binzhou Teachers College, 12(4) (1996), 32–34. (Chinese). [3] D. S. MITRINOVIĆ, Analytic Inequalities, Springer-Verlag, 1970. J. Inequal. Pure and Appl. Math., 3(4) Art. 56, 2002 http://jipam.vu.edu.au/