7.2
CiteScore
3.7
Impact Factor
Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors
Search in posts
Search in pages
Filter by Categories
ABUNDANCE ESTIMATION IN AN ARID ENVIRONMENT
Case Study
Editorial
Invited review
Letter to the Editor
Original Article
REVIEW
Review Article
SHORT COMMUNICATION
7.2
CiteScore
3.7
Impact Factor
Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors
Search in posts
Search in pages
Filter by Categories
ABUNDANCE ESTIMATION IN AN ARID ENVIRONMENT
Case Study
Editorial
Invited review
Letter to the Editor
Original Article
REVIEW
Review Article
SHORT COMMUNICATION
View/Download PDF

Translate this page into:

Short communication
30 (
4
); 549-551
doi:
10.1016/j.jksus.2017.05.010

On the Jensen functional and superterzaticity

Faculty of Sciences and Arts, Valahia University of Târgovişte, Bd. Unirii 18, Târgovişte, RO 130082, Romania
Disclaimer:
This article was originally published by Elsevier and was migrated to Scientific Scholar after the change of Publisher.

Peer review under responsibility of King Saud University.

Abstract

In this note we describe some results concerning upper and lower bounds for the Jensen functional. We use several known and new results to shed light on the concept of superterzatic function. Particular cases of interest are also considered.

Keywords

Jensen functional
Superterzatic functions
1

1 Introduction

The aim of this paper is to discuss new results concerning the Jensen functional in the framework of superterzatic functions.

For the reader’s convenience, before going into details, we quote here some relevant results regarding the superterzaticity and the Jensen functional.

Definition 1

We consider a real valued function f defined on an interval I , x 1 , x 2 , , x n I and p 1 , p 2 , , p n 0 , 1 with i = 1 n p i = 1 . The Jensen functional is defined by J f , p , x = i = 1 n p i f x i - f i = 1 n p i x i (see Dragomir, 2006).

Definition 2

A function f defined on an interval I = [ 0 , a ) is superterzatic, if for each x ¯ I there exists a real number C ( x ¯ ) such that J ( f , p , x ) i = 1 n p i x i ( x i - x ¯ ) C ( x ¯ ) + f ( | x i - x ¯ | ) | x i - x ¯ | = i = 1 n p i ( x i - x ¯ ) 2 C ( x ¯ ) + i = 1 n p i x i f ( | x i - x ¯ | ) | x i - x ¯ | for all x i 0 , i = 1 , , n , and p i 0 , i = 1 , , n , with i = 1 n p i = 1 , such that x ¯ = i = 1 n p i x i . We use the convention f ( 0 ) / 0 = 0 .

This definition was mentioned by S. Abramovich in her talk given at Conference on Inequalities and Applications 10, (Abramovich et al., 2012).

The set of superterzatic functions is closed under addition and positive scalar multiplication.

Example 1

(Abramovich et al., 2012) Let f x = x p , p 3 . This function is superterzatic with C ( x ¯ ) = p x ¯ p - 1 . The function g x = x 3 log 1 x is also subterzatic (i.e. the inequality in Definition 2 holds with reversed sign).

The paucity of literature on this topic motivated us to study the class of superterzatic functions and to emphasize some basic results connected to the Jensen functional and its behavior in the context of superterzaticity (See also Abramovich and Persson, 2013).

The Jensen functional has been already investigated under various assumptions: convexity, strong convexity, quasiconvexity, superquadraticity and so on. Therefore there is a wide literature about it, we only recommend several recent papers: (Abramovich, 2014, 2016; Dehghan, 2016; Franjić and Pecaric, 2015; Mitroi-Symeonidis and Minculete, 2016a,b; Moradi et al., in press). However there is no comparison between these estimates and we cannot indicate at this point the best (the sharpest) one. Throughout this paper we deal with the class of superterzatic functions which eventualy contains functions that also belong to other classes mentioned above and then the interested reader may find more convenient estimates using the same general technique.

2

2 Main results

We introduce in a natural way a more general functional.

Definition 3

Assume that we have a real valued function f defined on an interval I, the real numbers p ij , i = 1 , , k and j = 1 , , n i such that p ij > 0 , j = 1 n i p ij = 1 for all i = 1 , , k (we denote p i = p i 1 , p i 2 , , p in i ), x i = x i 1 , x i 2 , , x in i I n i for all i = 1 , , k and q = q 1 , , q k , q i > 0 such that i = 1 k q i = 1 . We define the generalized Jensen functional by J k f , p 1 , , p k , q , x 1 , , x k : = j 1 , , j k = 1 n 1 , , n k p 1 j 1 p kj k f i = 1 k q i x ij i - f i = 1 k q i j = 1 n i p ij x ij .

We notice that for k = 1 this definition is reduced to Definition 1.

For more results concerning Jensen’s functional the reader is referred to the papers (Mitroi, 2011; Mitroi, 2012).

Before announcing the main result, let us give the following lemma that describes the behaviour of the functional under the superterzaticity condition:

Lemma 1

Let f , p i , x i , q be as in Definition 3. If f is superterzatic then we have J k f , p 1 , , p k , q , x 1 , , x k j 1 , , j k = 1 n 1 , , n k p 1 j 1 p kj k i = 1 k q i x ij i i = 1 k q i x ij i - x ¯ C ( x ¯ ) + f i = 1 k q i x ij i - x ¯ i = 1 k q i x ij i - x ¯ , where x ¯ = i = 1 k q i j = 1 n i p ij x ij .

Proof

Since j 1 , , j k = 1 n 1 , , n k p 1 j 1 p kj k = 1 and x ¯ = j 1 , , j k = 1 n 1 , , n k p 1 j 1 p kj k i = 1 k q i x ij i , we use Definition 2 and the conclusion follows. □

Theorem 1

Let f , p i , x i and q be as in Definition 3 and the positive real numbers r ij , i = 1 , , k and j = 1 , , n i such that j = 1 n i r ij = 1 for all i = 1 , , k . We denote r i = r i 1 , r i 2 , , r in i for all i = 1 , , k , m = min 1 j 1 n 1 1 j k n k p 1 j 1 p kj k r 1 j 1 r kj k , M = max 1 j 1 n 1 1 j k n k p 1 j 1 p kj k r 1 j 1 r kj k . If f is superterzatic, then: J k f , p 1 , , p k , q , x 1 , , x k - m J k f , r 1 , , r k , q , x 1 , , x k j 1 , , j k = 1 n 1 , , n k p 1 j 1 p kj k - mr 1 j 1 r kj k i = 1 k q i x ij i × i = 1 k q i x ij i - x ¯ C x ¯ + f i = 1 k q i x ij i - x ¯ i = 1 k q i x ij i - x ¯ + m i = 1 k q i j = 1 n i r ij x ij × i = 1 k q i j = 1 n i r ij - p ij x ij C x ¯ + f x ˇ - x ¯ x ˇ - x ¯ and M J k f , r 1 , , r k , q , x 1 , , x k - J k f , p 1 , , p k , q , x 1 , , x k j 1 , , j k = 1 n 1 , , n k Mr 1 j 1 r kj k - p 1 j 1 p kj k i = 1 k q i x ij i × i = 1 k q i x ij i - x ˇ C x ˇ + f i = 1 k q i x ij i - x ˇ i = 1 k q i x ij i - x ˇ + i = 1 k q i j = 1 n i r ij x ij × i = 1 k q i j = 1 n i r ij - p ij x ij C x ˇ + f x ˇ - x ¯ x ˇ - x ¯ . where x ¯ = i = 1 k q i j = 1 n i p ij x ij and x ˇ = i = 1 k q i j = 1 n i r ij x ij .

Proof

The first inequality. Obviously J k f , p 1 , , p k , q , x 1 , , x k - m J k f , r 1 , , r k , q , x 1 , , x k = j 1 , , j k = 1 n 1 , , n k p 1 j 1 p kj k - mr 1 j 1 r kj k f i = 1 k q i x ij i + mf i = 1 k q i j = 1 n i r ij x ij - f i = 1 k q i j = 1 n i p ij x ij .

Since x ¯ = i = 1 k q i j = 1 n i p ij x ij = j 1 , , j k = 1 n 1 , , n k p 1 j 1 p kj k - mr 1 j 1 r kj k i = 1 k q i x ij i + m i = 1 k q i j = 1 n i r ij x ij and j 1 , , j k = 1 n 1 , , n k p 1 j 1 p kj k - mr 1 j 1 r kj k + m = 1 , the conclusion follows by Lemma 1.

The second inequality. One has M J k f , r 1 , , r k , q , x 1 , , x k - J k f , p 1 , , p k , q , x 1 , , x k = j 1 , , j k = 1 n 1 , , n k Mr 1 j 1 r kj k - p 1 j 1 p kj k f i = 1 k q i x ij i + f i = 1 k q i j = 1 n i p ij x ij - Mf i = 1 k q i j = 1 n i r ij x ij .

Since x ˇ = i = 1 k q i j = 1 n i r ij x ij = j 1 , , j k = 1 n 1 , , n k Mr 1 j 1 r kj k - p 1 j 1 p kj k M i = 1 k q i x ij i + 1 M i = 1 k q i j = 1 n i r ij x ij and j 1 , , j k = 1 n 1 , , n k Mr 1 j 1 r kj k - p 1 j 1 p kj k M + 1 M = 1 , we may apply again Lemma 1 and the conclusion follows. □

The particular case p 1 = = p k = p and x 1 = = x k = x is of interest.

Corollary 1

Let us consider x = x 1 , x 2 , , x n I n , p = p 1 , p 2 , , p n 0 , 1 n with i = 1 n p i = 1 , q = q 1 , q 2 , , q k such that q i > 0 , i = 1 k q i = 1 ( 1 k n ) and r = r 1 , r 2 , , r n 0 , 1 n with i = 1 n r i = 1 . We put m = min 1 i 1 , , i k n p i 1 p i k r i 1 r i k , M = max 1 i 1 , i k n p i 1 p i k r i 1 r i k . We define J k f , p , q , x i 1 , , i k = 1 n p i 1 p i k f j = 1 k q j x i j - f i = 1 n p i x i . If f is superterzatic, then J k f , p , q , x - m J k f , r , q , x i 1 , , i k = 1 n p i 1 p i k - mr i 1 r i k j = 1 k q j x i j × j = 1 k q j x i j - x ¯ C x ¯ + f j = 1 k q j x i j - x ¯ j = 1 k q j x i j - x ¯ + m i = 1 n r i x i j = 1 n r j - p j x j C x ¯ + f x ˇ - x ¯ x ˇ - x ¯ and M J k f , r , q , x - J k f , p , q , x i 1 , , i k = 1 n Mr i 1 r i k - p i 1 p i k j = 1 k q j x i j × j = 1 k q j x i j - x ˇ C x ˇ + f j = 1 k q j x i j - x ˇ j = 1 k q j x i j - x ˇ + i = 1 n r i x i j = 1 n r j - p j x j C x ˇ + f x ˇ - x ¯ x ˇ - x ¯ , where x ¯ = j = 1 n p j x j and x ˇ = j = 1 n r j x j .

For the particular case k = 1 we get:

Corollary 2

For i = 1 , , n , we consider x i I , p i > 0 with i = 1 n p i = 1 and r i > 0 with i = 1 n r i = 1 . We denote m = min i = 1 , , n p i r i , M = max i = 1 , , n p i r i .

If f is a superterzatic function, then we have: J f , p , x - m J f , r , x i = 1 n p i - mr i x i x i - x ¯ C x ¯ + f x i - x ¯ x i - x ¯ + m i = 1 n r i x i i = 1 n r i - p i x i C x ¯ + f x ˇ - x ¯ x ˇ - x ¯ and M J f , r , x - J f , p , x i = 1 n Mr i - p i x i x i - x ˇ C x ˇ + f x i - x ˇ x i - x ˇ + i = 1 n r i x i i = 1 n r i - p i x i C x ˇ + f x ˇ - x ¯ x ˇ - x ¯ , where x ¯ = j = 1 n p j x j and x ˇ = j = 1 n r j x j .

References

  1. , , , . On convex, superquadratic and superterzatic functions. In: , ed. Inequalities and Applications 2010, Internat. Ser. Numer. Math.. Vol 161. Basel: Birkhäuser; . p. :217-227.
    [Google Scholar]
  2. , , . Some new scales of refined Hardy type inequalities via functions related to superquadracity. Math. Inequal. Appl.. 2013;16(3):679-695.
    [Google Scholar]
  3. , . New applications of superquadracity. In: Analytic number theory, approximation theory, and special functions: In honor of Hari M. Srivastava. Springer; . pp. 365–395
    [Google Scholar]
  4. , . Applications of Quasiconvexity. In: Contributions in Mathematics and Engineering: In Honor of Constantin Carathéodory.. Springer; . pp 1–23
    [Google Scholar]
  5. , . Bounds for the normalised Jensen functional. Bull. Austral. Math. Soc.. 2006;74:471-478.
    [Google Scholar]
  6. , . Estimating the normalized Jensen functional. J. Math. Inequal. 2011;5(4):507-521.
    [Google Scholar]
  7. , . Connection between the Jensen and the Chebychev functionals. In: , ed. Inequalities and Applications 2010, Internat. Ser. Numer. Math.. Vol 161. Basel: Birkhä user; . p. :217-227.
    [CrossRef] [Google Scholar]
  8. , , . Refinements of the lower bounds of the jensen functional revisited, miskolc mathematical notes. Miskolc Math. Notes. 2015;16(2):833-842. ISSN 1787–2413
    [Google Scholar]
  9. , . Inequalities for the normalized Jensen functional with applications. Math. Slovaca. 2016;66(1):119-126.
    [CrossRef] [Google Scholar]
  10. , , . On the Jensen functional and superquadracity. Aequationes Math.. 2016;90(4):705-718.
    [CrossRef] [Google Scholar]
  11. , , . On the Jensen functional and the strong convexity. Bull. Malays. Math. Sci. Soc. 2016
    [CrossRef] [Google Scholar]
  12. H.R. Moradi, M.E. Omidvar, M.A. Khan, K. Nikodem, Around Jensen’s inequality for strongly convex functions, Aequationes mathematicae (in press).

Fulltext Views
8

PDF downloads
18
View/Download PDF
Download Citations
BibTeX
RIS
Show Sections