Gautschi's inequality

Last updated May 25, 2020

In real analysis, a branch of mathematics, Gautschi's inequality is an inequality for ratios of gamma functions. It is named after Walter Gautschi.

Statement

Let $x$ be a positive real number, and let $s \in (0, 1)$ . Then^[1]

x^{1-s}<{\frac {\Gamma (x+1)}{\Gamma (x+s)}}<(x+1)^{1-s}.

History

In 1948, Wendel proved the inequalities

\left({\frac {x}{x+s}}\right)^{1-s}\leq {\frac {\Gamma (x+s)}{x^{s}\Gamma (x)}}\leq 1

for $x > 0$ and $s \in (0, 1)$ .^[2] He used this to determine the asymptotic behavior of a ratio of gamma functions. The upper bound in this inequality is stronger than the one given above.

In 1959, Gautschi independently proved two inequalities for ratios of gamma functions. His lower bounds were identical to Wendel's. One of his upper bounds was the one given in the statement above, while the other one was sometimes stronger and sometimes weaker than Wendel's.

Consequences

An immediate consequence is the following description of the asymptotic behavior of ratios of gamma functions:

\lim _{x\to \infty }{\frac {\Gamma (x+1)}{\Gamma (x+s)x^{1-s}}}=1.

Proofs

There are several known proofs of Gautschi's inequality. One simple proof is based on the strict logarithmic convexity of Euler's gamma function. By definition, this means that for every $u$ and $v$ and every $t \in (0, 1)$ , we have

\Gamma (tu+(1-t)v)<\Gamma (u)^{t}\Gamma (v)^{1-t}.

Apply this inequality with $u = x$ , $v = x + 1$ , and $t = 1 - s$ . Also apply it with $u = x + s$ , $v = x + s + 1$ , and $t = s$ . The resulting inequalities are:

{\begin{aligned}\Gamma (x+s)&<\Gamma (x)^{1-s}\Gamma (x+1)^{s}=x^{s-1}\Gamma (x+1),\\\Gamma (x+1)&<\Gamma (x+s)^{s}\Gamma (x+s+1)^{1-s}=(x+s)^{1-s}\Gamma (x+s).\end{aligned}}

Rearranging the first of these gives the lower bound, while rearranging the second and applying the trivial estimate $x+s<x+1$ gives the upper bound.

Related inequalities

A survey of inequalities for ratios of gamma functions was written by Qi.^[3]

The proof by logarithmic convexity gives the stronger upper bound

{\frac {\Gamma (x+1)}{\Gamma (x+s)}}<(x+s)^{1-s}.

Gautschi's original paper proved a different stronger upper bound,

{\frac {\Gamma (x+1)}{\Gamma (x+s)}}\leq \exp((1-s)\psi (x+1)),

where $\psi$ is the digamma function. Neither of these upper bounds is always stronger than the other.^[4]

Kershaw proved two tighter inequalities. Again assuming that $x > 0$ and $s \in (0, 1)$ ,^[5]

{\begin{aligned}\left(x+{\frac {s}{2}}\right)^{1-s}&<{\frac {\Gamma (x+1)}{\Gamma (x+s)}}<\left[x-{\frac {1}{2}}+\left(s+{\frac {1}{4}}\right)^{1/2}\right]^{1-s},\\\exp \left((1-s)\psi (x+s^{1/2})\right)&<{\frac {\Gamma (x+1)}{\Gamma (x+s)}}<\exp \left((1-s)\psi \left(x+{\frac {1}{2}}(s+1)\right)\right).\end{aligned}}

Gautschi's inequality is specific to a quotient of gamma functions evaluated at two real numbers having a small difference. However, there are extensions to other situations. If $x$ and $y$ are positive real numbers, then the convexity of $\psi$ leads to the inequality:^[6]

{\frac {1}{2}}(\psi (x)+\psi (y))\leq {\frac {\log \Gamma (y)-\log \Gamma (x)}{y-x}}\leq \psi \left({\frac {x+y}{2}}\right).

For $s \in (0, 1)$ , this leads to the estimates

\exp {\bigl (}(1-s)\psi (x+s){\bigr )}\leq {\frac {\Gamma (x+1)}{\Gamma (x+s)}}\leq \exp \left((1-s)\psi \left(x+{\frac {1}{2}}(s+1)\right)\right).

A related but weaker inequality can be easily derived from the mean value theorem and the monotonicity of $\psi$ .^[7]

A more explicit inequality valid for a wider class of arguments is due to Kečkić and Vasić, who proved that if $y > x > 1$ , then:^[8]

{\frac {y^{y-1}}{x^{x-1}}}e^{x-y}<{\frac {\Gamma (y)}{\Gamma (x)}}<{\frac {y^{y-1/2}}{x^{x-1/2}}}e^{x-y}.

In particular, for $s \in (0, 1)$ , we have:

{\frac {(x+1)^{x}}{(x+s)^{x+s-1}}}e^{-(1-s)}<{\frac {\Gamma (x+1)}{\Gamma (x+s)}}<{\frac {(x+1)^{x+1/2}}{(x+s)^{x+s-1/2}}}e^{-(1-s)}.

Guo, Qi, and Srivastava proved a similar-looking inequality, valid for all $y > x > 0$ :^[9]

{\frac {(x+1)^{x+1}}{(y+1)^{y+1}}}e^{y-x}<{\frac {\Gamma (x+1)}{\Gamma (y+1)}}<{\frac {(x+1/2)^{x+1/2}}{(y+1/2)^{y+1/2}}}e^{y-x}.

For $s \in (0, 1)$ , this leads to:

{\frac {(x+1)^{x+1}}{(x+s)^{x+s}}}e^{s-1}<{\frac {\Gamma (x+1)}{\Gamma (x+s)}}<{\frac {(x+1/2)^{x+1/2}}{(x+s-1/2)^{x+s-1/2}}}e^{s-1}.

Related Research Articles

The Cauchy distribution, named after Augustin Cauchy, is a continuous probability distribution. It is also known, especially among physicists, as the Lorentz distribution, Cauchy–Lorentz distribution, Lorentz(ian) function, or Breit–Wigner distribution. The Cauchy distribution $is the distribution of the x -intercept of a ray issuing from with a uniformly distributed angle. It is also the distribution of the ratio of two independent normally distributed random variables with mean zero.$

In quantum mechanics, the uncertainty principle is any of a variety of mathematical inequalities asserting a fundamental limit to the precision with which the values for certain pairs of physical quantities of a particle, such as position, x, and momentum, p, can be predicted from initial conditions. Such variable pairs are known as complementary variables or canonically conjugate variables, and, depending on interpretation, the uncertainty principle limits to what extent such conjugate properties maintain their approximate meaning, as the mathematical framework of quantum physics does not support the notion of simultaneously well-defined conjugate properties expressed by a single value. The uncertainty principle implies that it is in general not possible to predict the value of a quantity with arbitrary certainty, even if all initial conditions are specified.

In probability theory, Chebyshev's inequality guarantees that, for a wide class of probability distributions, no more than a certain fraction of values can be more than a certain distance from the mean. Specifically, no more than 1/k² of the distribution's values can be more than k standard deviations away from the mean. The rule is often called Chebyshev's theorem, about the range of standard deviations around the mean, in statistics. The inequality has great utility because it can be applied to any probability distribution in which the mean and variance are defined. For example, it can be used to prove the weak law of large numbers.

In probability theory and statistics, the gamma distribution is a two-parameter family of continuous probability distributions. The exponential distribution, Erlang distribution, and chi-squared distribution are special cases of the gamma distribution. There are three different parametrizations in common use:

With a shape parameter k and a scale parameter θ.
With a shape parameter α = k and an inverse scale parameter β = 1/θ, called a rate parameter.
With a shape parameter k and a mean parameter μ = kθ = α/β.

In mathematics, the polygamma function of order $m$ is a meromorphic function on the complex numbers $ℂ$ defined as the $(m + 1)$ th derivative of the logarithm of the gamma function:

In mathematics, the digamma function is defined as the logarithmic derivative of the gamma function:

In number theory, the Mertens function is defined for all positive integers n as

In probability theory, the Chernoff bound, named after Herman Chernoff but due to Herman Rubin, gives exponentially decreasing bounds on tail distributions of sums of independent random variables. It is a sharper bound than the known first- or second-moment-based tail bounds such as Markov's inequality or Chebyshev's inequality, which only yield power-law bounds on tail decay. However, the Chernoff bound requires that the variates be independent – a condition that neither Markov's inequality nor Chebyshev's inequality require, although Chebyshev's inequality does require the variates to be pairwise independent.

Dirichlet distribution Probability distribution

In probability and statistics, the Dirichlet distribution, often denoted $, is a family of continuous multivariate probability distributions parameterized by a vector of positive reals. It is a multivariate generalization of the beta distribution, hence its alternative name of multivariate beta distribution (MBD) . Dirichlet distributions are commonly used as prior distributions in Bayesian statistics, and in fact the Dirichlet distribution is the conjugate prior of the categorical distribution and multinomial distribution.$

In probability theory, Hoeffding's inequality provides an upper bound on the probability that the sum of bounded independent random variables deviates from its expected value by more than a certain amount. Hoeffding's inequality was proven by Wassily Hoeffding in 1963.

The Wheeler–DeWitt equation is a field equation. It is part of a theory that attempts to combine mathematically the ideas of quantum mechanics and general relativity, a step towards a theory of quantum gravity. In this approach, time plays a role different from what it does in non-relativistic quantum mechanics, leading to the so-called 'problem of time'. More specifically, the equation describes the quantum version of the Hamiltonian constraint using metric variables. Its commutation relations with the diffeomorphism constraints generate the Bergman–Komar "group".

In probability theory and statistics, the inverse gamma distribution is a two-parameter family of continuous probability distributions on the positive real line, which is the distribution of the reciprocal of a variable distributed according to the gamma distribution. Perhaps the chief use of the inverse gamma distribution is in Bayesian statistics, where the distribution arises as the marginal posterior distribution for the unknown variance of a normal distribution, if an uninformative prior is used, and as an analytically tractable conjugate prior, if an informative prior is required.(Hoff, 2009:74)

In physics, the ARGUS distribution, named after the particle physics experiment ARGUS, is the probability distribution of the reconstructed invariant mass of a decayed particle candidate in continuum background.

In mathematics, in the area of complex analysis, Nachbin's theorem is commonly used to establish a bound on the growth rates for an analytic function. This article provides a brief review of growth rates, including the idea of a function of exponential type. Classification of growth rates based on type help provide a finer tool than big O or Landau notation, since a number of theorems about the analytic structure of the bounded function and its integral transforms can be stated. In particular, Nachbin's theorem may be used to give the domain of convergence of the generalized Borel transform, given below.

In the mathematical analysis, and especially in real and harmonic analysis, a Birnbaum–Orlicz space is a type of function space which generalizes the L^p spaces. Like the L^p spaces, they are Banach spaces. The spaces are named for Władysław Orlicz and Zygmunt William Birnbaum, who first defined them in 1931.

In geometry, a ball is a region in space comprising all points within a fixed distance from a given point; that is, it is the region enclosed by a sphere or hypersphere. An $n$ -ball is a ball in $n$ -dimensional Euclidean space. The volume of a unit $n$ -ball is an important expression that occurs in formulas throughout mathematics; it generalizes the notion of the volume enclosed by a sphere in 3-dimensional space.

In mathematics, the generalized polygamma function or balanced negapolygamma function is a function introduced by Olivier Espinosa Aldunate and Victor H. Moll.

The Kelvin–Stokes theorem, named after Lord Kelvin and George Stokes, also known as the Stokes' theorem, the fundamental theorem for curls or simply the curl theorem, is a theorem in vector calculus on $. Given a vector field, the theorem relates the integral of the curl of the vector field over some surface, to the line integral of the vector field around the boundary of the surface.$

In mathematics and physics, Lieb–Thirring inequalities provide an upper bound on the sums of powers of the negative eigenvalues of a Schrödinger operator in terms of integrals of the potential. They are named after E. H. Lieb and W. E. Thirring.

In mathematical physics, the Gordon decomposition of the Dirac current is a splitting of the charge or particle-number current into a part that arises from the motion of the center of mass of the particles and a part that arises from gradients of the spin density. It makes explicit use of the Dirac equation and so it applies only to "on-shell" solutions of the Dirac equation.

References

↑ NIST Digital Library of Mathematical Functions, 5.6.4.
↑ J.G. Wendel, Note on the Gamma function, Amer. Math. Monthly 55 (9) (1948) 563–564.
↑ Feng Qi, Bounds for the Ratio of Two Gamma Functions, Journal of Inequalities and Applications, Volume 2010, doi:10.1155/2010/493058.
↑ Feng Qi, Bounds for the ratio of two Gamma functions, J. Inequal. Appl. (2010) 1–84.
↑ D. Kershaw, Some extensions of W. Gautschi’s inequalities for the gamma function, Math. Comp. 41 (1983) 607–611.
↑ M. Merkle, Conditions for convexity of a derivative and applications to the Gamma and Digamma function, Facta Universitatis (Niš), Ser. Math. Inform. 16 (2001), 13-20.
↑ A. Laforgia, P. Natalini, Exponential, gamma and polygamma functions: Simple proofs of classical and new inequalities, J. Math. Anal. Appl. 407 (2013), 495–504.
↑ J. D. Kečkić and P. M. Vasić, Some inequalities for the gamma function, Publications de l’Institut Mathématique, vol. 11 (25), pp. 107–114, 1971.
↑ S. Guo, F. Qi, and H. M. Srivastava, Necessary and sufficient conditions for two classes of functions to be logarithmically completely monotonic, Integral Transforms and Special Functions, vol. 18, no. 11-12, pp. 819–826, 2007, https://dx.doi.org/10.1080/10652460701528933.

Gautschi Walter, (1959), Some Elementary Inequalities Relating to the Gamma and Incomplete Gamma Function, Journal of Mathematics and Physics, 38, doi:10.1002/sapm195938177.

This page is based on this Wikipedia article
Text is available under the CC BY-SA 4.0 license; additional terms may apply.
Images, videos and audio are available under their respective licenses.

[1] NIST Digital Library of Mathematical Functions, 5.6.4.

[2] J.G. Wendel, Note on the Gamma function, Amer. Math. Monthly 55 (9) (1948) 563–564.

[3] Feng Qi, Bounds for the Ratio of Two Gamma Functions, Journal of Inequalities and Applications, Volume 2010, doi:10.1155/2010/493058.

[4] Feng Qi, Bounds for the ratio of two Gamma functions, J. Inequal. Appl. (2010) 1–84.

[5] D. Kershaw, Some extensions of W. Gautschi’s inequalities for the gamma function, Math. Comp. 41 (1983) 607–611.

[6] M. Merkle, Conditions for convexity of a derivative and applications to the Gamma and Digamma function, Facta Universitatis (Niš), Ser. Math. Inform. 16 (2001), 13-20.

[7] A. Laforgia, P. Natalini, Exponential, gamma and polygamma functions: Simple proofs of classical and new inequalities, J. Math. Anal. Appl. 407 (2013), 495–504.

[8] J. D. Kečkić and P. M. Vasić, Some inequalities for the gamma function, Publications de l’Institut Mathématique, vol. 11 (25), pp. 107–114, 1971.

[9] S. Guo, F. Qi, and H. M. Srivastava, Necessary and sufficient conditions for two classes of functions to be logarithmically completely monotonic, Integral Transforms and Special Functions, vol. 18, no. 11-12, pp. 819–826, 2007, https://dx.doi.org/10.1080/10652460701528933.