Next Article in Journal
A Galerkin/POD Reduced-Order Model from Eigenfunctions of Non-Converged Time Evolution Solutions in a Convection Problem
Next Article in Special Issue
On the Converse Jensen-Type Inequality for Generalized f-Divergences and Zipf–Mandelbrot Law
Previous Article in Journal
A Formal Analysis of Generalized Peterson’s Syllogisms Related to Graded Peterson’s Cube
Previous Article in Special Issue
A Note of Jessen’s Inequality and Their Applications to Mean-Operators
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

An Application of Hayashi’s Inequality for Differentiable Functions

1
Department of Mathematics, Faculty of Science and Information Technology, Irbid National University, P.O. Box 2600, Irbid 21110, Jordan
2
Faculty of Science, University of Split, Ruđera Boškovića 33, 21000 Split, Croatia
*
Author to whom correspondence should be addressed.
Mathematics 2022, 10(6), 907; https://doi.org/10.3390/math10060907
Submission received: 1 March 2022 / Revised: 9 March 2022 / Accepted: 10 March 2022 / Published: 11 March 2022
(This article belongs to the Special Issue Advances in Mathematical Inequalities and Applications)

Abstract

:
In this work, we offer new applications of Hayashi’s inequality for differentiable functions by proving new error estimates of the Ostrowski- and trapezoid-type quadrature rules.

1. Introduction

In ([1], pp. 311–312) Hayashi proved the following theorem.
Theorem 1.
Let p : [ a , b ] R be a nonincreasing mapping on [ a , b ] and h : [ a , b ] R an integrable mapping on [ a , b ] with 0 h ( x ) A for all x [ a , b ] . Then, the inequality
A b λ b p ( x ) d x a b p ( x ) h ( x ) d x A a a + λ p ( x ) d x
holds, where λ = 1 A a b h ( x ) d x .
Inequality (1), called Hayashi’s inequality, is a simple generalization of Steffensen’s inequality which holds with same assumptions with A = 1 . For recent results concerning Hayashi’s inequality see [2].
In 1996, Agarwal and Dragomir [3] presented an application of this inequality as follows.
Theorem 2.
Let f : I R R be a differentiable mapping on I (the interior of I) and [ a , b ] I with M = sup x a , b f x < , m = inf x a , b f x < and M > m . If f is integrable on [ a , b ] , then the following inequality holds
1 b a a b f t d t f a + f b 2 f b f a m b a M b a f b + f a 2 M m m a M m b a 8 .
This elegant inequality presents an error estimation for the trapezoidal rule.
In 2002, Gauchman [4] generalized (2) for n-times differentiable functions using the Taylor expansion so that (2) becomes a special case of Gauchman’s result when n = 0 .
In this paper, we present a generalization of (2). In the same argument, two other inequalities of the Ostrowski and trapezoidal type are also introduced.

2. The Results

Let us begin with a generalization of (2).
Theorem 3.
Let g : [ a , b ] R be an absolutely continuous function on [ a , b ] with 0 g t b a , and suppose that g is integrable on [ a , b ] . Then
1 b a a b g t d t x a g a + b x g b b a λ x a + b 2 λ 2 b a λ b a 2 8
for all x [ a , b ] , where λ = g b g a b a . In particular, for x = a + b 2 , the following inequality holds
1 b a a b g t d t g a + g b 2 λ 2 b a λ b a 2 8 .
Proof. 
Let f t = x t , t [ a , b ] . Applying Hayashi’s inequality (1) by setting p t = f t and h t = g t , we get
b a b λ b ( x t ) d t a b ( x t ) g t d t b a a a + λ ( x t ) d t
where, A = b a or we write
λ = 1 b a a b g t d t = g b g a b a .
Also, we have
b λ b ( x t ) d t = λ x b + 1 2 λ 2 ,
a b ( x t ) g ( t ) d t = x a g a b x g b + a b g t d t ,
and
a a + λ ( x t ) d t = λ x a 1 2 λ 2 .
Substituting the above equalities in (4) and dividing by ( b a ) , we get
1 ( x ) : = λ x b + 1 2 λ 2 x a g a + b x g b b a + 1 b a a b g t d t : = I λ x a 1 2 λ 2 : = 2 ( x ) .
We also have
I 1 ( x ) + 2 ( x ) 2 = I λ x a + b 2 2 ( x ) 1 ( x ) 2 = λ 2 b a λ
which proves the first inequality in (3). The second inequality follows by applying the same technique as in ([3], pp. 96–97). □
Remark 1.
For some results closely related to Theorem 3 we refer the reader to [5,6,7,8,9,10,11,12,13].
A corrected generalized version of the Agarwal-Dragomir result (2) is incorporated in the following corollary.
Corollary 1.
Let g : [ a , b ] R be an absolutely continuous function on [ a , b ] with γ g t Γ , and suppose that g is integrable on [ a , b ] . Then
1 b a a b g t d t x a g a + b x g b b a g b g a b a · x a + b 2 Γ γ 2 · λ · b a λ b a Γ γ b a 8
for all x [ a , b ] , where λ = g b g a γ b a Γ γ . In particular, for x = a + b 2 , the following inequality holds
1 b a a b g t d t g a + g b 2 Γ γ 2 · λ · b a λ b a Γ γ b a 8 .
Proof. 
Repeating the proof of Theorem 3, with h ( t ) = g ( t ) γ , t [ a , b ] , we get the first inequality. The second inequality in (5) follows by applying the same technique as in ([3], pp. 96–97). Analogous manipulation for x = a + b 2 gives the same result as in (2). □
Remark 2.
Let the assumptions of Corollary 1 be satisfied. Then Corollaries 3 and 4 and Remarks 1 and 2 in [3] (p. 97) also hold.
In 1997, Dragomir and Wang [11] introduced an inequality of Ostrowski-Grüss type as follows: inequality
f x 1 b a a b f t d t f b f a b a x a + b 2 1 4 b a Γ γ
holds for all x [ a , b ] , where f is assumed to be differentiable on [ a , b ] with f L 1 [ a , b ] and γ f x Γ , x [ a , b ] .
In 1998, another result for twice differentiable was proved in [10]. In 2000, the constant 1 4 in (6) was improved by 1 3 in [14].
A better improvement of (6) can be deduced by applying Hayashi’s inequality as it is presented in the following theorem.
Theorem 4.
Let g : [ a , b ] R be an absolutely continuous function on [ a , b ] with 0 g t b a and suppose that g is integrable on [ a , b ] . Then
1 b a a b g t d t g x + λ x a + b 2 λ b a 2 λ 2 b a 2 16
for all x [ a , b ] , where λ = g b g a b a . In particular, for x = a + b 2 , the following inequality holds
1 b a a b g t d t g a + b 2 λ b a 2 λ 2 b a 2 16 .
Proof. 
Fix x [ a , b ] and let f t = a t , t [ a , x ] . Applying Hayashi’s inequality (1) by setting p t = f t and h t = g t , we get
b a x λ x ( a t ) d t a x ( a t ) g t d t b a a a + λ ( a t ) d t
where,
λ = 1 b a a b g t d t = g b g a b a .
Also, we have
x λ x ( a t ) d t = λ x a + 1 2 λ 2 ,
a x ( a t ) g ( t ) d t = x a g x + a x g t d t ,
and
a a + λ ( a t ) d t = 1 2 λ 2 .
Substituting in (8), we get
b a λ x a + 1 2 λ 2 x a g x + a x g t d t 1 2 λ 2 b a .
Now, let f t = b t , t [ x , b ] . Applying Hayashi’s inequality (1) again we get
b a b λ b ( b t ) d t x b ( b t ) g t d t b a x x + λ ( b t ) d t
where,
b λ b ( b t ) d t = 1 2 λ 2 ,
x b ( b t ) g ( t ) d t = b x g x + x b g t d t ,
and
x x + λ ( b t ) d t = λ b x 1 2 λ 2 .
Substituting in (10), we obtain
1 2 λ 2 b a b x g x + x b g t d t b a λ b x 1 2 λ 2 .
Adding (9) and (11) we get
b a λ x a + λ 2 a b g t d t b a g x b a λ b x λ 2 .
Setting
I : = 1 b a a b g t d t g x
1 ( x ) = λ x a + λ 2 ,
and
2 ( x ) = λ b x λ 2 .
Therefore,
I 1 ( x ) + 2 ( x ) 2 = 1 b a a b g t d t g x + λ x a + b 2 2 ( x ) 1 ( x ) 2 = λ b a 2 λ 2 ,
which proves the first inequality in (7). To prove the second inequality, define the mapping ϕ ( t ) = t 2 + b a 2 t . Then max ϕ ( t ) = ϕ b a 4 = b a 4 2 , so that ϕ ( λ ) = λ 2 + b a 2 λ b a 4 2 , which completes the proof of this theorem. □
Corollary 2.
Let g : [ a , b ] R be an absolutely continuous function on [ a , b ] with γ g t Γ and suppose that g is integrable on [ a , b ] . Then
1 b a a b g t d t g x + g b g a b a · x a + b 2 Γ γ b a λ b a 2 λ 2 b a Γ γ 16
for all x [ a , b ] , where λ = g b g a γ b a Γ γ . In particular, for x = a + b 2 , the following inequality holds
1 b a a b g t d t g a + b 2 Γ γ b a λ b a 2 λ 2 b a Γ γ 16 .
Proof. 
Repeating the proof of Theorem 4, with h ( t ) = g ( t ) γ , t [ a , b ] , we get the first inequality. The second inequality (12) follows by applying the same technique. □
Remark 3.
As we notice, (12) improves (6) by 1 4 , which is better than Matić et al. result from [14].
In [6], under the assumptions of Theorem 4, Alomari proved the following version of a Guessab–Schmeisser-type inequality (see [12]):
g x + g a + b x 2 1 b a a b g t d t Γ γ b a 8 ,
for all x a , a + b 2 .
Next we give an improvement of (13).
Theorem 5.
Let g : [ a , b ] R be an absolutely continuous function on [ a , b ] with 0 g t b a and suppose that g is integrable on [ a , b ] . Then
1 b a a b g t d t g x + g a + b x 2 λ b a 2 3 2 λ b a 2 24
for all x a , a + b 2 , where λ = g b g a b a .
Proof. 
Fix x a , a + b 2 and let f t = a t , t [ a , x ] . Applying Hayashi’s inequality (1) by setting p t = f t and h t = g t , we get that (8) holds, i.e.,
b a λ x a + 1 2 λ 2 x a g x + a x g t d t 1 2 λ 2 b a .
where
λ = 1 b a a b g t d t = g b g a b a .
Now, let f t = a + b 2 t , t x , a + b x . Applying Hayashi’s inequality (1) again we get
b a a + b x λ a + b x a + b 2 t d t x a + b x a + b 2 t g t d t b a x x + λ a + b 2 t d t
where
a + b x λ a + b x a + b 2 t d t = λ a + b 2 x + 1 2 λ 2 ,
x a + b x a + b 2 t g ( t ) d t = a + b 2 x g x + g a + b x + x a + b x g t d t ,
and
x x + λ a + b 2 t d t = λ a + b 2 x 1 2 λ 2 .
Substituting in (16), we get
b a λ a + b 2 x 1 2 λ 2 a + b 2 x g x + g a + b x + x a + b x g t d t b a λ a + b 2 x 1 2 λ 2 .
Now, let f t = b t , t [ a + b x , b ] . Applying Hayashi’s inequality (1) again we obtain
b a b λ b b t d t a + b x b b t g t d t b a a + b x a + b x + λ b t d t
where
b λ b ( b t ) d t = 1 2 λ 2 ,
a + b x b b t g ( t ) d t = x a g a + b x + a + b x b g t d t ,
and
a + b x a + b x + λ ( b t ) d t = λ x a 1 2 λ 2 .
Substituting in (18) we get
1 2 λ 2 b a x a g a + b x + a + b x b g t d t b a λ x a 1 2 λ 2 .
Adding (15), (17) and (19) we obtain
λ b a b a 2 3 2 λ a b g t d t b a g x + g a + b x 2 λ b a b a 2 3 2 λ ,
which is equivalent to the first inequality in (14). To prove the second inequality in (14), define the mapping ϕ ( t ) = 3 2 b a t 2 + b a 2 2 t . Then max ϕ ( t ) = ϕ b a 6 = b a 2 24 , so that ϕ λ = 3 2 b a λ 2 + b a 2 2 λ b a 2 24 , which completes the proof. □
A generalization of (13) and (14) is incorporated in the following result.
Corollary 3.
Let g : [ a , b ] R be an absolutely continuous function on [ a , b ] with γ g t Γ and suppose that g is integrable on [ a , b ] . Then
1 b a a b g t d t g x + g a + b x 2 1 2 λ Γ γ b a b a 3 λ Γ γ b a 24
for all x a , a + b 2 , where λ = g b g a γ b a Γ γ .
Proof. 
Repeating the proof of Theorem 5, with h ( t ) = g ( t ) γ , t [ a , b ] , we get the first inequality. The second inequality in (20) follows by applying the same technique. □

3. Applications

Let X be a random variable taking values in the finite interval [ a , b ] , with the probability density function f : [ a , b ] [ 0 , 1 ] with the cumulative distribution function F ( x ) = P r ( X x ) = a b f ( t ) d t .
Theorem 6.
With the assumptions of Theorem 4, we have the inequality
1 2 F x + F a + b x b E X b a 1 2 λ Γ γ b a b a 3 λ Γ γ b a 24
for all x a , a + b 2 , where λ = F b F a γ b a Γ γ , and E ( X ) is the expectation of X.
Proof. 
In the proof of Corollary 3, we set f = F , and take into account that
E X = a b t d F t = b a b F t d t .
We leave the details to the interested reader. □

4. Conclusions

This paper summarises several types of general quadrature rules, such as the general trapezoidal rule or the so-called Ostrowski trapezoidal, Ostrowski midpoint and Guessab–Schmeisser quadrature rules for symmetric points. Using the presented inequalities, several error estimates of the above quadrature rules could therefore be derived with corresponding numerical experiments. This work thus represents a very good application of Hayashi’s inequality in quadrature approximation. One future research direction might be to use Fink’s generalization of the Ostrowski inequality to obtain some Hayashi–Ostrowski-type results. Further applications of Hayashi’s inequality we leave to the interested reader.

Author Contributions

Conceptualization, M.W.A.; Formal analysis, M.W.A. and M.K.B.; Funding acquisition, M.K.B.; Investigation, M.K.B.; Methodology, M.W.A.;Writing original draft preparation, M.W.A. and M.K.B. All authors have read and agreed to the published version of the manuscript.

Funding

This publication was supported by the University of Split, Faculty of Science.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Mitrinović, D.S.; Pexcxarixcx, J.E.; Fink, A.M. Classical and New Inequalities in Analysis. In Mathematics and Its Applications; East European Series; Kluwer Academic Publishers Group: Dordrecht, The Netherlands, 1993; Volume 61. [Google Scholar]
  2. Alomari, M.W.; Hussain, S.; Liu, Z. Some Steffensen’s type inequalities. Adv. Pure Appl. Math. 2017, 8, 219–226. [Google Scholar] [CrossRef]
  3. Agarwal, R.P.; Dragomir, S.S. An application of Hayashi’s inequality for differentiable functions. Comput. Math. Appl. 1996, 32, 95–99. [Google Scholar] [CrossRef] [Green Version]
  4. Gauchman, H. Some integral inequalities involving Taylor’s remainder I. J. Inequal. Pure Appl. Math. 2002, 3, 26. [Google Scholar]
  5. Alomari, M.W. A companion of the generalized trapezoid inequality and applications. J. Math. Appl. 2013, 36, 5–15. [Google Scholar]
  6. Alomari, M.W. A companion of Ostrowski’s inequality for mappings whose first derivatives are bounded and applications in numerical integration. Kragujev. J. Math. 2012, 36, 77–82. [Google Scholar]
  7. Alomari, M.W. New inequalities of Steffensen’s type for s–convex functions. Afr. Mat. 2014, 25, 1053–1062. [Google Scholar] [CrossRef]
  8. Alomari, M.W. A companion of Dragomir’s generalization of Ostrowski’s inequality and applications in numerical integration. Ukr. Math. J. 2012, 64, 491–510. [Google Scholar] [CrossRef] [Green Version]
  9. Cerone, P.; Dragomir, S.S.; Pearce, C.E.M. A generalized trapezoid inequality for functions of bounded variation. Turk. J. Math. 2000, 24, 147–163. [Google Scholar]
  10. Cerone, P.; Dragomir, S.S.; Roumeliotis, J. An inequality of Ostrowski-Griiss type for twice differentiable mappings and applications in numerical integration. RGMIA Res. Rep. Collect. 1998, 1, 8. [Google Scholar]
  11. Dragomir, S.S.; Wang, S. An inequality of Ostrowski–Grüss’ type and its applications to the estimation of error bounds for some special means and for Some numerical quadrature rules. Comput. Math. Appl. 1997, 33, 15–20. [Google Scholar] [CrossRef] [Green Version]
  12. Guessab, A.; Schmeisser, G. Sharp integral inequalities of the Hermite-Hadamard type. J. Approx. Theory 2002, 115, 260–288. [Google Scholar] [CrossRef] [Green Version]
  13. Ujević, N. New bounds for the first inequality of Ostrowski–Grüss type and applications. Comput. Math. Appl. 2003, 46, 421–427. [Google Scholar] [CrossRef] [Green Version]
  14. Matić, M.; Pexcxarixcx, J.; Ujevixcx, N. Improvement and further generalization of inequalities of Ostrowski–Grüss type. Comput. Math. Appl. 2000, 39, 161–175. [Google Scholar] [CrossRef] [Green Version]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Alomari, M.W.; Klaričić Bakula, M. An Application of Hayashi’s Inequality for Differentiable Functions. Mathematics 2022, 10, 907. https://doi.org/10.3390/math10060907

AMA Style

Alomari MW, Klaričić Bakula M. An Application of Hayashi’s Inequality for Differentiable Functions. Mathematics. 2022; 10(6):907. https://doi.org/10.3390/math10060907

Chicago/Turabian Style

Alomari, Mohammad W., and Milica Klaričić Bakula. 2022. "An Application of Hayashi’s Inequality for Differentiable Functions" Mathematics 10, no. 6: 907. https://doi.org/10.3390/math10060907

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop