- Research
- Open access
- Published:
Next generalization of Cîrtoaje’s inequality
Journal of Inequalities and Applications volume 2017, Article number: 159 (2017)
Abstract
In this paper, we classify sets of solutions of the next generalized Cîrtoaje’s inequality and its reverse, respectively.
1 Introduction
In recent years, inequalities with power-exponential functions have been intensively studied [1–7]. They have many important applications. For example, they can be found in mathematical analysis and in other theories like mathematical physics, mathematical biology, ordinary differential equations, probability theory and statistics, chemistry, economics. For more details, a literature review and the history of inequalities with power-exponential functions, see [2]. Cîrtoaje, in [1], has introduced the following interesting conjecture on the inequalities with power-exponential functions. The inequality is similar to the reverse arithmetic-geometric mean inequality where its terms were rearranged.
Conjecture 1
If \(a,b \in(0,1]\) and \(r \in(0,e]\), then
The conjecture was proved by Matejíčka [3]. Matejíčka also proved (1) under other conditions in [4, 5]. For example, he showed that (1) is also valid for \(a,b,r \in (0,e]\). In [5], one interesting property of the generalized Cîrtoaje’s inequality (CI) was found. In [6], a classification of sets of solutions of (CI)
was made.
2 Methods
In this paper, methods of mathematical and numerical analysis are used. We make a classification of sets of solutions of the other generalization of (CI).
Let φ, ψ be functions from \(\{1,\ldots,n\}\) to \(\{1,\ldots,n\}\), where \(n\in N\). Put
The function \(F(r)\) is defined on \(R^{n}_{+}\) where \(n\in\mathbf{N}\), \(r\geq0\), \(R^{n}_{+}=\{(x_{1},\ldots,x_{n}), x_{i}>0, i=1,\ldots,n\}\). We note that \(F(r)\geq0\) is equivalent to the following generalization of Cîrtoaje’s inequality (I):
The reverse inequality to (I)
we denote by (RI).
3 Results and discussion
We remark that in [6] the special case of our classification for (4) was presented, where \(\varphi(i)=i\), \(\psi(i)=i+1\), \(i=1,\ldots,n-1\), \(\varphi(n)=n\), \(\psi(n)=1\).
The following functions:
we will call characteristic functions of (I).
Put
We prove the following lemma.
Lemma 1
Let \(F(r)\) be defined by (3). Let φ, ψ be arbitrary functions from \(\{1,\ldots,n\}\) to \(\{ 1,\ldots,n\}\), \(n\in N\). Then \(F(r)\) is a concave function for each \(A\in R^{n}_{+}-S^{n}\), and \(F(0)=0\). If there is \(i\neq j\); \(i,j\in N\) such that \(x_{\psi(i)}\ln x_{i}\neq x_{\psi(j)}\ln x_{j}\), then \(F(r)\) is a strongly concave function in A.
Proof
\(F(0)=0\) is evident. Easy calculation gives
and
where
The proof is completed. □
Now we prove the following lemma.
Lemma 2
Let g, h be defined by (6), (7). Let φ, ψ be arbitrary functions from \(\{1,\ldots,n\}\) to \(\{1,\ldots,n\}\), \(n\in N\). Then there are five cases.
-
1.
If \(h(x_{1},\ldots,x_{n})=\sum_{i=1}^{n} (x_{\varphi (i)}-x_{\psi(i)} )\log(x_{i})<0\) for \(A=(x_{1},\ldots,x_{n})\in R^{n}_{+}\) then (RI) is valid for all \(r>0\) in \(A=(x_{1},\ldots,x_{n})\in R^{n}_{+}\).
-
2.
If \(h(x_{1},\ldots,x_{n})=\sum_{i=1}^{n} (x_{\varphi (i)}-x_{\psi(i)} )\log(x_{i})=0\) and \(g(x_{1},\ldots,x_{n})= \frac{1}{n}\sum_{i=1}^{n}x_{\varphi(i)}\log (x_{i})-\max_{1\leq m\leq n}\{x_{\psi(m)}\log(x_{m})\}<0\) for \(A=(x_{1},\ldots,x_{n})\in R^{n}_{+}\) then (RI) is valid for all \(r>0\) in \(A=(x_{1},\ldots,x_{n})\in R^{n}_{+}\).
-
3.
If \(h(x_{1},\ldots,x_{n})=\sum_{i=1}^{n} (x_{\varphi (i)}-x_{\psi(i)} )\log(x_{i})=0\) and \(g(x_{1},\ldots,x_{n})= \frac{1}{n}\sum_{i=1}^{n}x_{\varphi(i)}\log (x_{i})-\max_{1\leq m\leq n}\{x_{\psi(m)}\log(x_{m})\}=0\) for \(A=(x_{1},\ldots,x_{n})\in R^{n}_{+}\) then \(F(r)=0\) for \(r\geq0\) in \(A=(x_{1},\ldots,x_{n})\in R^{n}_{+}\).
-
4.
If \(h(x_{1},\ldots,x_{n})=\sum_{i=1}^{n} (x_{\varphi (i)}-x_{\psi(i)} )\log(x_{i})>0\) and \(g(x_{1},\ldots,x_{n})= \frac{1}{n}\sum_{i=1}^{n}x_{\varphi(i)}\log (x_{i})-\max_{1\leq m\leq n}\{x_{\psi(m)}\log(x_{m})\}\geq0\) for \(A=(x_{1},\ldots,x_{n})\in R^{n}_{+}\) then (I) is valid for all \(r\geq0\) in \(A=(x_{1},\ldots,x_{n})\in R^{n}_{+}\).
-
5.
If \(h(x_{1},\ldots,x_{n})=\sum_{i=1}^{n} (x_{\varphi (i)}-x_{\psi(i)} )\log(x_{i})>0\) and \(g(x_{1},\ldots,x_{n})= \frac{1}{n}\sum_{i=1}^{n}x_{\varphi(i)}\log (x_{i})-\max_{1\leq m\leq n}\{x_{\psi(m)}\log(x_{m})\}<0\) for \(A=(x_{1},\ldots,x_{n})\in R^{n}_{+}\) then there is \(r_{0}>0\) such that (I)is valid for \(r\in(0,r_{0}]\) and (RI) is valid for \(r\in(r_{0},\infty)\) in \(A=(x_{1},\ldots,x_{n})\in R^{n}_{+}\).
Proof
The proof is evident. It follows from Lemma 1. □
Note 1
It is easy to see that if \(g(x_{1},\ldots,x_{n})=0\) and \(h(x_{1},\ldots,x_{n})=0\) then \(F(r)=0\) for all \(r\geq0\). Really, from Lemma 1 we have \(F'(0)=0\) and \(\lim_{r\rightarrow\infty }F'(r)=0\). If \(F(r_{1})\neq0\) for some \(r_{1}>0\) then \(F(r_{1})<0\). Then there exists z such that \(F(r_{1})-F(0)=F'(z)r_{1}\) and \(0< z< r_{1}\). It implies \(F'(z)<0\). Because of \(F''(r)\leq0\) we get \(F'\) is non-increasing for \(r\geq0\). For \(r>z>0\) we obtain \(F'(r)\leq F'(z)\) so \(\lim_{r\rightarrow\infty}F'(r)\leq F'(z)<0\). This is a contradiction.
4 Conclusion
In this paper, we showed the following. If (I) is valid in \((x_{1},\ldots,x_{n})\) for some \(r_{0}>0\) then (I) is valid in \((x_{1},\ldots,x_{n})\) for all \(0< r\leq r_{0}\). Similarly, if (RI) is valid in \((x_{1},\ldots,x_{n})\) for some \(r_{0}>0\) then (RI) is valid in \((x_{1},\ldots,x_{n})\) for all \(r>r_{0}\).
We think that the way how to classify sets of solutions of the power-exponential inequalities could be used for other suitable inequalities.
Now we give examples of concrete applications of our results. We make the complete classification of sets of solutions for (I) and (RI) inequalities where \(n=2\). Using Matlab for plotting graphs of the solution curves for the characteristic equations \(g(X)=0\), \(h(X)=0\) we obtain the following figures for (I) and (RI). In the figures we denote by \(I+\mathit{RI}\) the points where (I) and also (RI) are locally valid. By I we denote points where (I) is valid for all \(r>0\) and by RI we denote points where (RI) is valid for all \(r>0\).
It is easy to show that for \(n=2\) there are only 12 basic cases of inequalities (I). The other four cases of (I) can be transformed to the previous cases.
Example 1
Let \(n=2\), \(\varphi(1)=1\), \(\varphi(2)=2\), \(\psi(1)=1\), \(\psi(2)=2\). Then we have (I):
See Figure 1.
Example 2
Let us consider that \(n=2\), \(\varphi(1)=2\), \(\varphi(2)=1\), \(\psi (1)=1\), \(\psi(2)=2\). Then by (I)
See Figure 2.
Example 3
Put \(n=2\), \(\varphi(1)=1\), \(\varphi(2)=2\), \(\psi(1)=1\), \(\psi(2)=1\). Then we obtain by (I)
See Figure 3.
Example 4
Let us consider \(n=2\), \(\varphi(1)=2\), \(\varphi(2)=1\), \(\psi(1)=1\), \(\psi(2)=1\). Then we get by (I)
See Figure 4.
Example 5
Let \(n=2\), \(\varphi(1)=1\), \(\varphi(2)=1\), \(\psi(1)=2\), \(\psi(2)=2\). Then we have by (I)
See Figure 5.
Example 6
Put \(n=2\), \(\varphi(1)=1\), \(\varphi(2)=2\), \(\psi(1)=2\), \(\psi(2)=1\). Then by (I)
See Figure 6.
Example 7
Let us consider \(n=2\), \(\varphi(1)=1\), \(\varphi(2)=1\), \(\psi(1)=1\), \(\psi(2)=1\). Then we obtain by (I)
See Figure 1.
Example 8
Put \(n=2\), \(\varphi(1)=2\), \(\varphi(2)=2\), \(\psi(1)=2\), \(\psi(2)=2\). Then by (I)
See Figure 1.
Example 9
Let us consider \(n=2\), \(\varphi(1)=1\), \(\varphi(2)=1\), \(\psi(1)=1\), \(\psi(2)=2\). Then we get by (I)
See Figure 7.
Example 10
Let \(n=2\), \(\varphi(1)=1\), \(\varphi(2)=1\), \(\psi(1)=2\), \(\psi(2)=1\). Then by (I)
See Figure 8.
Example 11
Put \(n=2\), \(\varphi(1)=2\), \(\varphi(2)=1\), \(\psi(1)=2\), \(\psi(2)=1\). We obtain by (I)
See Figure 1.
Example 12
Let \(n=2\), \(\varphi(1)=2\), \(\varphi(2)=2\), \(\psi(1)=2\), \(\psi(2)=1\). Then we have by (I)
See Figure 9.
References
Cîrtoaje, V: Proofs of three open inequalities with power exponential functions. J. Nonlinear Sci. Appl. 4(2), 130-137 (2011)
Coronel, A, Huancas, F: The proof of three power exponential inequalities. J. Inequal. Appl. 2014, 509 (2014). http://www.journalofinequalitiesandapplications.com/content/2014/1/509
Matejíčka, L: Proof of one open inequality. J. Nonlinear Sci. Appl. 7, 51-62 (2014). http://www.tjnsa.com
Matejíčka, L: On the Cîrtoaje’s conjecture. J. Inequal. Appl. 2016, 152 (2016). doi:10.1186/s13660-016-1092-2
Matejíčka, L: Some remarks on Cîrtoaje’s conjecture. J. Inequal. Appl. 2016, 269 (2016). doi:10.1186/s13660-016-1211-0
Matejíčka, L: On one interesting inequality. Surv. Math. Appl. 11, 1842-6298 (2016)
Miyagi, M, Nishizawa, Y: A stronger inequality of Cîrtoaje’s one with power exponential functions. J. Nonlinear Sci. Appl. 8, 224-230 (2015). http://www.tjnsa.com
Acknowledgements
The work was supported by VEGA grant no. 1/0649/17 and KEGA grant no. 007TnUAD-4/2017. The author thanks Professor Ondrušová, Dean of the faculty FPT TnUAD in Púchov, Professor Vavro and Ing. Balážová for their kind support and he is deeply grateful to the unknown reviewer for valuable remarks and suggestions.
Author information
Authors and Affiliations
Corresponding author
Additional information
Dedicated to my wife Emília Matejíčková
Competing interests
The author declares that he has no competing interests.
Author’s contributions
All authors read and approved the final manuscript.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
About this article
Cite this article
Matejíčka, L. Next generalization of Cîrtoaje’s inequality. J Inequal Appl 2017, 159 (2017). https://doi.org/10.1186/s13660-017-1436-6
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s13660-017-1436-6