Abstract
In this paper, we focus on the generalized Marcum function of the second kind of order \(\nu >0\), defined by
where \(a>0, b\ge 0,\) \(K_{\nu }\) stands for the modified Bessel function of the second kind, and \(c_{a,\nu }\) is a constant depending on a and \(\nu \) such that \(R_{\nu }(a,0)=1.\) Our aim is to find some new tight bounds for the generalized Marcum function of the second kind and compare them with the existing bounds. In order to deduce these bounds, we include the monotonicity properties of various functions containing modified Bessel functions of the second kind as our main tools. Moreover, we demonstrate that our bounds in some sense are the best possible ones.
Similar content being viewed by others
1 Introduction and preliminary results
Let \(R_{\nu }(a,b)\) be the so-called generalized Marcum function of the second kind [4], defined by
where \(a>0, b\ge 0\), \(\nu >0\), and \(K_{\nu }\) stands for the modified Bessel function of the second kind. As \(R_{\nu }(a,0)=1\), we can write \(R_{\nu }(a,b)\) as follows:
where \(a>0,\) \(b\ge 0\), and \(\nu >0\). In [4], Baricz et al. proved some monotonicity properties of the generalized Marcum function of the second kind with respect to a, b and \(\nu \), and log-concavity and convexity properties with respect to the parameter b. They also deduced several bounds for the Marcum function of the second kind (that is, when \(\nu =1\)) and for the generalized Marcum function of the second kind.
The study of the generalized Marcum function of the second kind is motivated by the importance of the generalized Marcum Q-function in the electrical engineering literature, which was studied by several authors in the last few decades, see for example [1, 2, 9] and the references therein. In particular, Baricz and Sun [2] obtained several tight bounds for the generalized Marcum Q-function which are based on the monotonicity properties of the functions of the form \(x\mapsto x^{\alpha \nu +\beta }I_{\nu }(x)/f(e^x,\sinh x,\cosh x),\) where \(\alpha \) and \(\beta \) are real numbers and \(I_{\nu }\) stands for the modified Bessel function of the first kind. These bounds are sharp in nature in the sense that they cannot be further improved. Motivated by the work of Baricz and Sun [2], in this paper our aim is to find some new bounds for the generalized Marcum function of the second kind. These bounds are obtained from the monotonicity properties of the functions of the form \(x\mapsto x^{\alpha \nu +\beta }K_{\nu }(x)/f(e^x,\sinh x,\cosh x),\) where \(\alpha \) and \(\beta \) are real numbers. The monotonicity properties of these functions depend upon some monotonicity properties of the ratio of modified Bessel functions of the second kind, which were investigated in [5] and [11]. The bounds which we deduce in this paper are tight enough and cannot be further improved as we prove that the power \(\alpha \nu +\beta \) is the smallest or largest constant so that the corresponding monotonicity property for the function \(x\mapsto x^{\alpha \nu +\beta }K_{\nu }(x)/f(e^x,\sinh x,\cosh x)\) holds.
It is worth mentioning that during the preparation of our manuscript we found out that the distribution whose survival function we study in this paper resembles a distribution of Nadarajah [6]. More precisely, motivated by a Bayesian inference of an inverse Gaussian sample, Nadarajah [6] introduced a new modified Bessel distribution of the second kind supported on the real line. In [4], it has been shown that the generalized Marcum function of the second kind is in fact the survival function of the truncated distribution of a special case (when \(p=1/2\)) of the modified Bessel distribution of the second kind considered by Nadarajah. Taking into account the vast application field in engineering and physical problems of the generalized Marcum Q-function and of the Bessel distributions, we believe that our paper may be useful for people who meet in their study such kind of exponential integrals involving the modified Bessel function of the second kind.
This paper is organized as follows: in this section we prove the preliminary results which contain the monotonicity of the functions involving the modified Bessel function of the second kind. Using the preliminary results in Sect. 2, we present some new lower and upper bounds for the generalized Marcum function of the second kind for two cases: when \(b\ge a>0\) and \(a>b>0\). In Sect. 3, we provide some comparison between our new bounds which are obtained in Sect. 2 and the bounds given in [4] for both cases when \(b\ge a>0\) and \(a>b>0\). The discussion is concluded in Sect. 4, by illustrating the theoretical results numerically.
It is interesting to note that the functions \(x\mapsto x^{\nu }K_{\nu }(x)e^x\) and \(x\mapsto x^{-\nu }K_{\nu }(x)e^x\) have different monotonic behavior, see [4]. More specifically
-
(1)
for all \(\nu \ge 1/2\), the function \(x\mapsto x^{\nu }K_{\nu }(x)e^x\) is increasing on \((0,\infty )\).
-
(2)
for all \(\nu \ge -1/2\), the function \(x\mapsto x^{-\nu }K_{\nu }(x)e^x\) is decreasing on \((0,\infty )\).
Moreover, Yang and Zheng [11] showed that for fixed \(\nu \ge 1/2\) the function \(x\mapsto \sqrt{x}K_{\nu }(x)e^x\) is decreasing on \((0,\infty )\). Then the natural question which we can ask is that for a fixed \(\nu \) what is the monotonic behavior of the function \(x\mapsto x^{\alpha _{\nu }}K_{\nu }(x)e^x\) on \((0,\infty )\)? Does there exist the smallest \(\alpha _{\nu }\) such that the function \(x\mapsto x^{\alpha _{\nu }}K_{\nu }(x)e^x\) is increasing on \((0,\infty )\)? Does there exist the largest \(\alpha _{\nu }\) such that the function \(x\mapsto x^{\alpha _{\nu }}K_{\nu }(x)e^x\) is decreasing on \((0,\infty )\)? These questions will be answered in this section. Moreover, this study improves one of the results of Yang and Zheng [11] and provides an alternative proof for [10, Corollary 1]. Similarly, the behavior of the functions \(x\mapsto x^{\beta _{\nu }}K_{\nu }(x)/ \sinh x \) and \(x\mapsto x^{\beta _{\nu }}K_{\nu }(x)/ \cosh x\) are studied and we find the largest and the smallest powers of x such that these functions possess the monotonic decreasing/increasing property. This section also discusses the monotonic behavior of the functions \(x\mapsto x^{\delta _{\nu }}K_{\nu }(x) \sinh x \) and \(x\mapsto x^{\beta _{\nu }}K_{\nu }(x) \cosh x \).
The following two lemmas are used to prove the main Lemma in this section. For Lemma 1 we refer [11, Theorem 2.6], while for Lemma 2 we refer to [3, Theorem 2] and [11, Lemma 2.4].
Lemma 1
The function \(\nu \mapsto {K'_{\nu }(x)}/{K_{\nu }(x)}\) is strictly decreasing on \((0,\infty )\) for \(x>0\).
Lemma 2
The function \(x \mapsto {xK'_{\nu }(x)}/{K_{\nu }(x)}\) is decreasing on \((0,\infty )\) for all \(\nu \in {\mathbb {R}}\).
The main Lemma in this section is the following result.
Lemma 3
Let \(\nu \in {\mathbb {R}}\). Then the following assertions are true:
- a.:
-
For \(\nu \ge 1/2\) the function \(x\mapsto \sqrt{x} K_{\nu }(x)e^x\) is strictly decreasing on \((0,\infty )\).
- b.:
-
For \(\nu >1/2\) the smallest constant \(\alpha _{\nu }\) for which the function \(x\mapsto x^{\alpha _{\nu }} K_{\nu }(x)e^x\) is increasing on \((0,\infty )\) is \(\alpha _{\nu }= \nu \). For \(\nu \ge 1/2\) the largest constant \(\alpha _{\nu }\) for which the function \(x\mapsto x^{\alpha _{\nu }} K_{\nu }(x)e^x\) is decreasing on \((0,\infty )\) is \(\alpha _{\nu }= 1/2\).
- c.:
-
The largest constant \(\beta _{\nu }\) for which the function \(x\mapsto x^{\beta _{\nu }}K_{\nu }(x)/ \sinh x\) is strictly decreasing on \((0,\infty )\) is \(\beta _{\nu }=\nu +1\).
- d.:
-
The function \(x\mapsto x^{\nu }K_{\nu }(x)/\cosh x\) is strictly decreasing on \((0,\infty )\). Moreover, the largest constant \(\gamma _{\nu }\) for which the function \(x\mapsto x^{\gamma _{\nu }}K_{\nu }(x)/ \cosh x\) is strictly decreasing on \((0,\infty )\) is \(\gamma _{\nu }=\nu \).
- e.:
-
For \(\nu >1/2\) the function \(x\mapsto x^{\delta _{\nu }}K_{\nu }(x)\sinh x\) is increasing on \((0,\infty )\) when \(\delta _{\nu }\ge \nu \).
- f.:
-
For \(\nu >1/2\) the function \(x\mapsto x^{\epsilon _{\nu }}K_{\nu }(x)\cosh x\) is increasing on \((0,\infty )\) when \(\epsilon _{\nu }\ge \nu +\theta ,\) where
$$\begin{aligned} \theta =\sup _{x>0}\left( \frac{2x}{e^{2x}+1}\right) = 0.278{\ldots }. \end{aligned}$$For \(\nu \ge 1/2\) the largest constant for which the function \(x\mapsto x^{\epsilon _{\nu }}K_{\nu }(x)\cosh x\) is decreasing on \((0,\infty )\) is \(\epsilon _{\nu }=1/2\).
Proof
a. This follows from [11, Corollary 3.2].
b. Consider the function \(f_{\nu }:(0,\infty )\rightarrow {\mathbb {R}},\) defined by \(f_{\nu }(x)=x^{\alpha _{\nu }}K_{\nu }(x)e^x\). Then
In view of [7, 10.29.2]
we have
where \(\phi _{\nu }: (0,\infty )\rightarrow {\mathbb {R}}\) is defined by \(\phi _{\nu }(x)=x-{xK_{\nu -1}(x)}/{K_{\nu }(x)}\).
To find the smallest value of \(\alpha _{\nu }\) for which the function \(f_{\nu }\) is increasing on \((0,\infty )\), it is necessary to find the minimum of the function \(\phi _{\nu }\). By Soni inequality [8, Eq. 5] and the fact \(K_{\nu }(x)=K_{-\nu }(x)\), we have
This gives
Now, in view of the asymptotic formula [7, 10.30.2]
where \(\nu >0\) and \(x\rightarrow 0,\) we get
By using (1.6) and (1.8) we obtain \(f'_{\nu }(x)\ge 0\) for \(\alpha _{\nu }\ge \nu \). Thus, \(\alpha _{\nu }=\nu \) is the smallest \(\alpha _{\nu }\) for which \(x\mapsto x^{\alpha _{\nu }}K_{\nu }(x)e^x\) is increasing on \((0,\infty )\).
Now, our aim is to find the largest constant \(\alpha _{\nu }\) for which the function \(x\mapsto x^{\alpha _{\nu }}K_{\nu }(x)e^x\) is decreasing on \((0,\infty )\) for all \(\nu \ge 1/2\). We again rewrite the derivative of \(f_{\nu }(x)\) as follows:
where \(\psi _{\nu }:(0,\infty )\rightarrow {\mathbb {R}}\) is defined by \(\psi _{\nu }(x)=1/2+x+xK'_{\nu }(x)/K_{\nu }(x)\). Thus, to find the largest constant for which the function \(f_{\nu }\) is decreasing it is necessary to find the maximum value of the function \(\psi _{\nu }\). Due to Lemma 1, we have that \(xK'_{\nu }(x)/K_{\nu }(x)\le xK'_{1/2}(x)/K_{1/2}(x)\) for any \(x>0\) and \(\nu \ge 1/2\). Now, by the recurrence relation (1.3) and using the fact that \(K_{\nu }(x)=K_{-\nu }(x)\) we get
Thus for all \(x>0\) we get
By using once again the recurrence relation (1.3) and in view of the asymptotic expansion [7, 10.40.2]
where
we conclude that
By using (1.11) and (1.12), it follows that \(f'(x)\le 0\) for all \(x\in (0,\infty )\) when \(\alpha _{\nu }\le 1/2\). Thus \(\alpha _{\nu }=1/2\) is the largest value for which the function \(x\mapsto e^{\alpha _{\nu }}e^x K_{\nu }(x)\) is decreasing on \((0,\infty )\).
c. Consider the function \(g_{\nu }:(0,\infty )\rightarrow {\mathbb {R}},\) defined by \(g_{\nu }(x)={x^{\beta _{\nu }}K_{\nu }(x)}/{\sinh x}\). Then
From Lemma 2 and the monotone decreasing property of \(x\mapsto -x \coth x\), we see that for all \(\nu \in {\mathbb {R}}\), the function \(x \mapsto xK'_{\nu }(x)/K_{\nu }(x)-x \coth x\) is a decreasing function on \((0,\infty )\). Consequently, the maximum value of \(xK'_{\nu }(x)/K_{\nu }(x)-x \coth x\) is nothing but \(\lim _{x\rightarrow 0}\left( xK'_{\nu }(x)/K_{\nu }(x)-x \coth x \right) \). Using the asymptotic formula (1.7) we get
Thus from (1.13) we obtain
and consequently for all \(\beta _{\nu }\le \nu +1\) we have that \(g'_{\nu }(x)\le 0\) for all \(x>0\). Hence \(x\mapsto x^{\beta _{\nu }}K_{\nu }(x)/\sinh x\) is a decreasing function for all \(\beta _{\nu }\le \nu +1\). Consequently \(\beta _{\nu }=\nu +1\) is the largest value for which the function \(x\mapsto x^{\beta _{\nu }}K_{\nu }(x)/ \sinh x\) is decreasing on \((0,\infty )\).
d. Observe that
For all \(\nu \in {\mathbb {R}}\) proceeding like in part c, the function \(x\mapsto xK'_{\nu }(x)/K_{\nu }(x) - x \tanh x\) is decreasing on \((0,\infty )\). Hence
Thus (1.14) becomes
Hence for all \(\gamma _{\nu }\le \nu \) the function \(x\mapsto x^{\gamma _{\nu }}K_{\nu }(x)/ \cosh x\) is decreasing on \((0,\infty )\). Consequently \(\gamma _{\nu }=\nu \) is the largest value for which the function \(x\mapsto x^{\gamma _{\nu }}K_{\nu }(x)/ \cosh x\) is decreasing on \((0,\infty )\).
e. First note that
From (1.5) and \(\coth x>1\), we have
Consequently for \(\nu >1/2\) and \(x>0\) we have
Hence for \(x>0\) we get \((x^{\delta _{\nu }}K_{\nu }(x) \sinh x)'\ge 0\) when \(\delta _{\nu }\ge \nu \). Thus \(x\mapsto x^{\delta _{\nu }} K_{\nu }(x) \sinh x\) is an increasing function on \((0,\infty )\) for all \(\delta _{\nu }\ge \nu \).
f. Note that
where
and the inequality (1.17) follows from (1.5). It is easy to check that the function \(\phi \) has maximum value at \(x_0=0.63923\ldots \) which equals \(\sup _{x>0}(\phi (x))=\phi (x_0)=0.278\ldots \). It then follows
On the other hand, by using the inequality (1.10), we have
where \(\phi (x)\) is defined in (1.18). Since \(\inf _{x>0} \phi (x)=0\), it gives
In view of
one can conclude that for \(\nu \ge 1/2\) the largest constant for which the function \(x\mapsto x^{\epsilon _{\nu }}K_{\nu }(x) \cosh x\) is decreasing on \((0,\infty )\) is \(\epsilon _{\nu }=1/2\). \(\square \)
Remark 1
Recall that the functions given in part c of [4, Lemma 5] and part a of Lemma 3 are the sharpest in terms of their monotonicity. From part b we can say that these functions cannot be further improved by raising or diminishing the power of x in the functions. As a consequence of part b of Lemma 3, we can conclude that if \(1/2<\alpha <\nu \), then the function \(x\mapsto x^{\alpha _{\nu }}e^x K_{\nu }(x)\) is neither monotonic increasing nor monotonic decreasing in \((0,\infty )\). Equivalently the function \(x\mapsto \alpha _{\nu } +x+xK'{\nu }(x)/K_{\nu }(x)\) has at least one real positive zero. For example, using the expression \(K_{3/2}(x)=\sqrt{(\pi /2)}(e^{-x}/\sqrt{x})(1+1/x)\), it is easy to verify that the function \(x\mapsto x e^x K_{3/2}(x)\) is decreasing on (0, 1) and increasing on \([1,\infty )\). In fact, by [10, Corollary 1], there is an \(x_0>0\) such that the function \(x\mapsto x^{\alpha _{\nu }}e^x K_{\nu }(x)\) is decreasing on \((0,x_0)\) and increasing on \((x_0,\infty )\).
Remark 2
It is worth mentioning that recently Yang and Tian [10, Corollary 1] obtained the result of part b of Lemma 3 by studying the properties of ratios of two Laplace transforms. In this paper an independent proof is provided.
Remark 3
As \( xK'_{\nu }(x)/K_{\nu }(x)-x\coth x \rightarrow -\infty \) whenever \(x\rightarrow \infty \), by using (1.13), we conclude that the function \(x\mapsto x^{\beta _{\nu }}K_{\nu }(x)/\sinh x\) is not an increasing function on the whole interval \((0,\infty )\) for any \(\beta _{\nu }\). Similarly, \(x\mapsto x^{\gamma _{\nu }}K_{\nu }(x)/\cosh x\) is not an increasing function on the whole interval \((0,\infty )\) for any \(\gamma _{\nu }\).
2 Lower and upper bounds for the generalized Marcum function of the second kind
In this section some new tight lower and upper bounds for the generalized Marcum function of the second kind are obtained by using the monotonicity property of the functions discussed in Lemma 3. The bounds are expressed in terms of the elementary special function named as the complementary error function \({\text {erfc}}(t)\). It is defined as follows [7, 7.2.2]
Recall also that for \(a, \nu >0\), \(c_{a,\nu }\) denotes the constant defined by
2.1 Case \(1: b\ge a>0\)
Theorem 1
For \(b\ge a>0\), the following inequalities hold:
where \(\nu \ge 3/2\) in (2.1), and \(\nu >0\) in (2.2) and (2.3).
Proof
For \(\nu \ge {1}/{2}\) and \(t \ge b\) part a of Lemma 3 leads to
which in view of (1.1) implies that
To prove (2.2), the monotone property of \(x\mapsto x^{\nu +1} K_{\nu }(x)/ \sinh x\) is used. Note that part c of Lemma 3 implies
for all \(t\ge b\) and \(\nu \in {\mathbb {R}},\) and this implies that
Similarly, from part d of Lemma (3) we get
for all \(t\ge b\) and \(\nu \in {\mathbb {R}}.\) By using the above inequality in view of (1.1) we get
\(\square \)
Theorem 2
For \(\nu >3/2\) and \(b\ge a>0\), the following inequalities hold true:
Proof
Using part e of Lemma 3 with constant \(\delta _{\nu } =\nu \) and proceeding like in the previous theorem we can deduce inequality (2.7). Similarly, part f of Lemma 3 with \(\epsilon _{\nu }=\nu +1\) yields the inequality (2.8) since part f of Lemma 3 is true for all \(\epsilon _{\nu }\ge \nu +0.278\ldots \). The choice \(\epsilon _{\nu }=\nu +1\) is just sake of convenience. \(\square \)
Remark 4
Using the well-known inequality \(\frac{1}{1-e^{-2at}}> 1\) for all \(a, t>0\), we can obtain from the lower bound (2.7) a weaker lower bound, which can be expressed in terms of the complementary error function as follows:
The best choice for \(\epsilon _{\nu }\) in part f of Lemma 3 is \(\epsilon _{\nu }=\nu +\theta \) where \(\theta =0.278\ldots \) which gives a tighter lower bound for \(R_{\nu }(a,b)\) than the lower bound given in (2.8), that is precisely
2.2 Case \(2: a> b>0\)
Theorem 3
For \(a>b>0\), the following inequalities hold:
where \(\nu \ge 3/2\) in (2.10) and \(\nu >0\) in (2.11) and (2.12).
Proof
For \(\nu >1/2\) and \(0<t\le b\), part a of Lemma 3 implies
which in view of (1.2) implies that
The inequality (2.11) is obtained by using the monotonicity of the function \(t\mapsto {t^{\nu +1}K_{\nu }(t)}/{\sinh t}\). For \(0<t\le b\) and \(\nu \in {\mathbb {R}}\), part c of Lemma 3 leads to
Using the above inequality in view of (1.2), for \(\nu >0\) we get
Hence
Similarly, by using part d of Lemma (3) we obtain the upper bound for \(R_{\nu }(a,b)\) given in (2.12). \(\square \)
Theorem 4
For \(\nu >3/2\) and \(a> b>0\), the following inequalities hold:
Proof
Using part e of Lemma 3 with constant \(\delta _{\nu } =\nu \) and proceeding similarly as in the previous theorem we can obtain inequality (2.15). Similarly, part f of Lemma 3 with \(\epsilon _{\nu }=\nu +1\) yields the inequality (2.16) since part f of Lemma 3 is true for all \(\epsilon _{\nu }\ge \nu +0.278\ldots \). The choice \(\epsilon _{\nu }=\nu +1\) is just sake of convenience. \(\square \)
Remark 5
The best choice for \(\epsilon _{\nu }\) in part f of Lemma 3 is \(\epsilon _{\nu }=\nu +\theta \) where \(\theta =0.278\ldots \) which gives a tighter lower bound for \(R_{\nu }(a,b)\) than the lower bound given in (2.16), that is precisely
For \(\nu >3/2\), we can get from the lower bound (2.16) for \(R_{\nu }(a,b)\) a weaker lower bound, which can be expressed in terms of complementary error function as follows:
3 Sharpness of the bounds and comparison with other existing bounds
In this section, we discuss the tightness of the bounds obtained in Sect. 2 and compare these bounds with the bounds given in [4]. All the bounds for \(R_{\nu }(a,b)\) stated in the previous section and the section 2 of [4] are obtained by using the bounds for \(K_{\nu }(t)\). Thus we compare the bounds for \(K_{\nu }(t),\) which give immediately the comparison of the bounds for \(R_{\nu }(a,b)\).
3.1 Comparison with other existing bounds
Let \(b\ge a>0.\) For \(\nu \ge 1/2\), Baricz et al. [4] obtained the following upper bound for \(R_{\nu }(a,b)\)
by using the inequality
In this paper, by using the inequality
for \(\nu \ge 1/2,\) a new upper bound is obtained for \(R_{\nu }(a,b)\) in the case when \(\nu \ge 3/2\)
On the other hand, for \(\nu >3/2\) we have
Hence the new bound (3.2) is sharper than the upper bound (3.1) obtained by Baricz et al. [4] for \(\nu >3/2\). For \(\nu >0\), Baricz et al. [4] obtained the following upper bound for \(R_{\nu }(a,b)\)
by using the monotone decreasing property of the function \(t\mapsto t^{-\nu }e^{-t}K_{\nu }(t)\) on \((0,\infty )\) and for \(t\ge b\) the inequality
Note that \(t\mapsto e^{-t}\cosh t\) and \(t\mapsto t\coth t\) are decreasing and increasing functions, respectively, on \((0,\infty )\). Thus for \(t\ge b>0\)
Consequently for \(t\ge b\) and \(\nu >0\)
In view of (3.5), (2.5), and (2.6), we conclude that the upper bound in (2.2) is sharper than the upper bound in (2.3). Moreover, in view of (3.4) and (3.5), we conclude that the new bound (2.3) is sharper than the upper bound in (3.3) obtained by Baricz et al. [4].
Remark 6
In view of (3.5) for \(\nu >0\) and \(t\ge b>0\), we get
By using this we get the following new upper bound for \(R_{\nu }(a,b)\)
In view of (3.5), we conclude that the new upper bound (3.6) is sharper than the upper bound (3.3) obtained by Baricz et al. [4].
For \(\nu \ge 3/2\), Baricz et al. [4] obtained the following lower bound for \(R_{\nu }(a,b)\)
by using the monotone increasing property of the function \(t\mapsto e^t t^{\nu } K_{\nu }(t)\) on \((0,\infty )\) for all \(\nu \ge 1/2\), that is, for \(t\ge b>0\) and \(\nu \ge 1/2\) the inequality
Note that \(t\mapsto e^{-t} \sinh t\) and \(t\mapsto t \coth t\) both are increasing functions on \((0,\infty )\). Thus for \(t\ge b>0\)
Consequently,
In view of (3.9), the new lower bound in (2.7) is sharper than the new lower bound in (2.8). Moreover, in view of (3.9) and (3.8), we conclude that the lower bound in (3.7) obtained by Baricz et al. [4] is sharper than the new lower bound in(2.7). It is easy to see that the lower bound in (2.9) is sharper than the lower bound in (2.8) but weaker than the lower bound in (3.7) which is obtained by Baricz et al. [4].
Similarly to the previous discussion, for the case when \(a>b>0\) we can verify the followings:
- i.:
-
The upper bound in (2.11) is sharper than the upper bound in (2.12), and the upper bound in (2.12) is sharper than the upper bound in (3.10) obtained by Baricz et al. [4]
$$\begin{aligned} R_{\nu }(a,b)\le 1-\frac{c_{a,\nu }}{(ab)^{\nu -1}} \frac{K_{\nu -1}(ab)}{e^{ab}} \int _{-a}^{b-a} (y+a)^{2\nu -1}e^{-\frac{y^2}{2}} \mathrm{d}y \text { for } \nu >0. \end{aligned}$$(3.10) - ii.:
-
The upper bound in (2.10) is sharper than the upper bound in (3.11) given in [4]
$$\begin{aligned} R_{\nu }(a,b)\le&1-\frac{c_{a,\nu }}{(ab)^{\nu -1}} K_{\nu -1}(ab) e^{ab} \nonumber \\&\times \int _{a}^{b+a} (y-a)^{2\nu -1}e^{-\frac{y^2}{2}} \mathrm{d}y \quad \text { for } \nu >\frac{1}{2}. \end{aligned}$$(3.11) - iii.:
-
The lower bound in (2.15) is sharper than the lower bound in (2.16), but weaker than the lower bound in (3.12) obtained by Baricz et al. [4]
$$\begin{aligned} R_{\nu }(a,b)&\ge 1- c_{a,\nu }\left( \frac{b}{a}\right) ^{\nu -1} e^{ab} K_{\nu -1}(ab)\nonumber \\&\quad \times \left[ e^{-\frac{a^2}{2}}-e^{-\frac{(b+a)^2}{2}}- a\sqrt{\frac{\pi }{2}} \left( {\text {erfc}}\left( \frac{a}{\sqrt{2}}\right) \nonumber \right. \right. \\&\left. \left. -{\text {erfc}}\left( \frac{b+a}{\sqrt{2}}\right) \right) \right] . \end{aligned}$$(3.12) - iv.:
-
The lower bound in (2.17) is sharper than the lower bound in (2.16) but weaker than the lower bound in (3.12).
Remark 7
By using the monotone decreasing property of the functions \(t\mapsto t^{-\nu }e^{-t}K_{\nu }(t)\), \(t\mapsto e^{-t}\cosh t\) and the monotone increasing property of the function \(t\mapsto t\cosh t\) we get
By comparing the first and the last one, we get the following new upper bound for \(\nu >0\),
Clearly, the new upper bound in (3.14) is sharper than the upper bound in (3.10) obtained by Baricz et al. [4].
3.2 Tightness of the bounds for \(b\rightarrow 0\) and \(b\rightarrow \infty \)
In this subsection, the tightness of the new upper and lower bounds are discussed. Section 3.1 shows that all the new upper bounds obtained in this study are sharper than the upper bounds obtained by Baricz et al. [4]. Moreover, the upper bounds obtained by Baricz et al. [4] are tight bounds. Consequently, all the new upper bounds obtained in this study are also tight. It is also possible to verify that the relative error for the upper bounds tend to zero as b approaches infinity. In this subsection, proofs are discussed to show that all the new lower bounds are tight as well as the relative error in the lower bounds tends to zero as \(b\rightarrow \infty \).
The following limits show the tightness of the relative error for the lower bounds:
and
Using the limit (3.15), we conclude that the lower bound \(L_2\) in (2.8) tends to zero as \(b\rightarrow \infty \). Moreover, by using (3.16) and the following asymptotic relation
we can verify that the relative error \((L_2-R_{\nu }(a,b))/R_{\nu }(a,b)\) tends to zero as \(b\rightarrow \infty \).
As the lower bound \(L_1\) in (2.7) is sharper than the lower bound \(L_2\) in (2.8), \(L_1\) is also tight as \(b\rightarrow \infty \). Similarly, we can verify that the relative error in \(L_1\) approaches to zero as \(b\rightarrow \infty \). Similarly, for the case \(a>b>0\) all the new lower bounds (2.15) and (2.16) are also tight as well as the relative error tends to zero as \(b\rightarrow 0\).
4 Numerical results
In this section, we present numerical results in order to compare the various upper and lower bounds of \(R_{\nu }(a,b)\). To draw the figures for the bounds of \(R_{\nu }(a,b)\) we have used Mathematica 8.0.
4.1 Case \(1: b\ge a>0\)
Let the upper bounds given in (2.1), (2.2), (2.3), (3.3), and (3.1) be denoted by \(U_1\), \(U_2\), \(U_3\), \(U_4\), and \(U_5\), respectively. Let \(L_1, L_2, L_3\), and \(L_4\) denote the lower bounds (2.7), (2.8), (3.7) and [4, Eq. 2.22].
For the choice \(\nu =3\) and \(a=1\), Fig. 1 illustrates the comparison between all the upper bounds with \(R_{3}(1,b)\) in the interval (1, 2.5). Let \(R_3\) denote \(R_{3}(1,b)\). Figure 1 supports the theoretical results obtained in Sect. 3. Note that the new upper bound \(U_1\) is sharper than all the other upper bounds. For large b, the upper bound \(U_5\) is sharper than the upper bound \(U_3\).
For the choice \(\nu =4\) and \(a=1\), Fig. 2 illustrates the comparison between all the upper bounds with \(R_{4}(1,b)\) in the interval (1, 2.5). Let \(R_4\) denote \(R_{4}(1,b)\). Figure 2 supports the theoretical results obtained in Sect. 3. Clearly, both lower bounds \(L_1\) and \(L_2\) are sharper than the lower bound \(L_4\).
4.2 Case \(2: a>b>0\)
Let the upper bounds given in (2.10), (2.11), (2.12), (3.10), and (3.11) be denoted by \(U_6\), \(U_7\), \(U_8\), \(U_9\), and \(U_{10},\) respectively. Let \(L_1, L_2, L_3\), and \(L_4\) denote the lower bounds (2.15), (2.16), (3.12), and [4, Eq. 2.23].
For the choice \(\nu =2\) and \(a=3\), Fig. 3 illustrates the comparison between all the upper bounds for \(R_{2}(3,b)\) in the interval (0, 3). Let \(R_2\) denote \(R_{2}(3,b)\). Figure 3 supports the theoretical results obtained in Sect. 3. Clearly, the new upper bounds \(U_7\) and \(U_8\) are better than the upper bound \(U_{10}\) for smaller range of b but for large values of b the upper bound \(U_{10}\) is better than the upper bounds \(U_7\) and \(U_8\).
For the choice \(\nu =3\) and \(a=1\), Fig. 4 illustrates the comparison between all the lower bounds with \(R_{3}(1,b)\) in the interval (0, 1). Let \(R_3\) denote \(R_{3}(1,b)\). Figure 4 supports the theoretical results obtained in Sect. 3.
References
Baricz, Á., Sun, Y.: New bounds for the generalized Marcum \(Q\)-function. IEEE Trans. Inform. Theory 55(7), 3091–3100 (2009)
Baricz, Á., Sun, Y.: Bounds for the generalized Marcum \(Q\)-function. Appl. Math. Comput. 217(5), 2238–2250 (2010)
Baricz, Á., Ponnusamy, S., Vuorinen, M.: Functional inequalities for modified Bessel functions. Expo Math. 29(4), 399–414 (2011)
Baricz, Á., Bisht, N., Singh, S., Vijesh, A.: The generalized Marcum function of the second kind: monotonicity patterns and tight bounds. J. Comput. Appl. Math. 382 (2021). Art. 113093
Ismail, M.E.H., Muldoon, M.E.: Monotonicity of the zeros of a cross-product of Bessel functions. SIAM J. Math. Anal. 9(4), 759–767 (1978)
Nadarajah, S.: A modified Bessel distribution of the second kind. Statistica 47(4), 405–413 (2007)
Olver, F.W.J., Lozier, D.W., Boisvert, R.F., Clark, C.W. (eds.): NIST Handbook of Mathematical Functions. Cambridge University Press, Cambridge (2010). (http://dlmf.nist.gov/)
Soni, R.P.: On an inequality for modified Bessel functions. J. Math. Phys. 44, 406–407 (1965)
Sun, Y., Baricz, Á.: Inequalities for the generalized Marcum \(Q\)-function. Appl. Math. Comput. 203(1), 134–141 (2008)
Yang, Z., Tian, J.F.: Monotonicity rules for the ratio of two Laplace transforms with applications. J. Math. Anal. Appl. 470, 821–845 (2019)
Yang, Z.H., Zheng, S.Z.: The monotonicity and convexity for the ratios of modified Bessel functions of the second kind and applications. Proc. Am. Math. Soc. 145(7), 2943–2958 (2017)
Funding
Open access funding provided by Óbuda University.
Author information
Authors and Affiliations
Corresponding author
Additional information
Dedicated to the memory of Richard Askey.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors are very thankful to the reviewers for a careful evaluation of the manuscript and their several helpful comments which improved the quality of the paper. N. Bisht is grateful to the Council of Scientific and Industrial Research India (Grant No. 09/1022(0014)/2013-EMR-I) for the financial support, and S. Singh is thankful to the Science and Engineering Research Board (SERB), Department of Science and Technology, Government of India for the financial support through Project CRG/2020/002875.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Baricz, Á., Bisht, N., Singh, S. et al. Bounds for the generalized Marcum function of the second kind. Ramanujan J 58, 1–21 (2022). https://doi.org/10.1007/s11139-021-00440-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11139-021-00440-9
Keywords
- Modified Bessel function of the second kind
- Bounds
- Survival function
- Complementary error function
- Generalized Marcum Q-function