Stable polynomial: Difference between revisions

From formulasearchengine
Jump to navigation Jump to search
→Properties: hurwitz stable and schur stable should be separate bullet points
 
oops typo criterium->criterion
Line 1: Line 1:
My name is Leonard Sanor. I life in Creil (France).<br><br>my blog; [http://support.legendarymarketing.com/entries/52708280-Dog-Training-Kinds-Of-Containment-Fence dog training in chicago]
In [[mathematics]], the '''Barnes G-function''' ''G''(''z'') is a [[function (mathematics)|function]] that is an extension of [[superfactorial]]s to the [[complex number]]s. It is related to the [[Gamma function]], the [[K-function]] and the [[Glaisher–Kinkelin constant]], and was named after [[mathematician]] [[Ernest William Barnes]].<ref>E.W.Barnes, "The theory of the G-function", ''Quarterly Journ. Pure and Appl. Math.'' '''31''' (1900), 264–314.</ref> Up to elementary factors, it is a special case of the [[double gamma function]].
 
 
Formally, the Barnes ''G''-function is defined in the following [[Weierstrass product]] form:
 
 
:<math> G(1+z)=(2\pi)^{z/2} \text{exp}\left(- \frac{z+z^2(1+\gamma)}{2} \right) \, \prod_{k=1}^{\infty}\left(1+\frac{z}{k}\right)^k \text{exp}\left(\frac{z^2}{2k}-z\right) </math>
 
 
where <math>\, \gamma \, </math> is the [[Euler–Mascheroni constant]], exp(''x'') = ''e''<sup>''x''</sup>, and ∏ is [[capital pi notation]].
 
 
==Functional equation and integer arguments==
 
 
The Barnes ''G''-function satisfies the [[functional equation]]
 
 
:<math> G(z+1)=\Gamma(z)\, G(z) </math>
 
 
with normalisation ''G''(1)&nbsp;=&nbsp;1. Note the similarity between the functional equation of the Barnes G-function and that of the Euler [[Gamma function]]:
 
 
:<math> \Gamma(z+1)=z \, \Gamma(z) </math>
 
 
The functional equation implies that G takes the following values at [[integer]] arguments:
 
 
:<math>G(n)=\begin{cases} 0&\text{if }n=0,-1,-2,\dots\\ \prod_{i=0}^{n-2} i!&\text{if }n=1,2,\dots\end{cases}</math>
 
 
and thus
 
:<math>G(n)=\frac{(\Gamma(n))^{n-1}}{K(n)}</math>
 
 
where <math>\,\Gamma(x)\,</math> denotes the [[Gamma function]] and ''K'' denotes the [[K-function]]. The functional equation uniquely defines the G function if the convexity condition: <math>\, \frac{d^3}{dx^3}G(x)\geq 0\, </math> is added.<ref>M. F. Vignéras, ''L'équation fonctionelle de la fonction zêta de Selberg du groupe mudulaire SL<math>(2,\mathbb{Z})</math>'', Astérisque '''61''', 235–249 (1979).</ref>
 
==Reflection formula 1.0==
 
The [[difference equation]] for the G function, in conjunction with the [[functional equation]] for the [[Gamma function]], can be used to obtain the following [[reflection formula]] for the Barnes G function (originally proved by [[Hermann Kinkelin]]):
 
 
:<math> \log G(1-z) = \log G(1+z)-z\log 2\pi+ \int_0^z \pi x \cot \pi x \, dx.</math>
 
 
The logtangent integral on the right-hand side can be evaluated in terms of the [[Clausen function]] (of order 2), as is shown below:
 
 
:<math>2\pi \log\left( \frac{G(1-z)}{G(1+z)} \right)= 2\pi z\log\left(\frac{\sin\pi z}{\pi} \right)+\text{Cl}_2(2\pi z)</math>
 
 
The proof of this result hinges on the following evaluation of the cotangent integral: introducing the notation <math>\, Lc(z)\, </math> for the logtangent integral, and using the fact that <math>\,(d/dx) \log(\sin\pi x)=\pi\cot\pi x\,</math>, an integration by parts gives
 
 
:<math>Lc(z)=\int_0^z\pi x\log(\sin \pi x)\,dx=z\log(\sin \pi z)-\int_0^z\log(\sin \pi x)\,dx=</math>
 
 
:<math>z\log(\sin \pi z)-\int_0^z\Bigg[\log(2\sin \pi x)-\log 2\Bigg]\,dx=</math>
 
 
:<math>z\log(2\sin \pi z)-\int_0^z\log(2\sin \pi x)\,dx</math>
 
 
Performing the integral substitution <math>\, y=2\pi x \Rightarrow dx=dy/(2\pi)\,</math> gives
 
 
:<math>z\log(2\sin \pi z)-\frac{1}{2\pi}\int_0^{2\pi z}\log\left(2\sin \pi \frac{y}{2} \right)\,dy</math>
 
 
The [[Clausen function]] - of second order - has the integral representation
 
 
:<math>\text{Cl}_2(\theta) = -\int_0^{\theta}\log\Bigg|2\sin \frac{x}{2} \Bigg|\,dx</math>
 
 
However, within the interval <math>\, 0 < \theta < 2\pi \,</math>, the [[absolute value]] sign within the [[integrand]] can be omitted, since within the range the 'half-sine' function in the integral is strictly positive, and strictly non-zero. Comparing this definition with the result above for the logtangent itegral, the following relation clearly holds:
 
 
:<math>Lc(z)=z\log(2\sin \pi z)+\frac{1}{2\pi}\, \text{Cl}_2(2\pi z)</math>
 
 
Thus, after a slight rearrangement of terms, the proof is complete:
 
 
:<math>2\pi \log\left( \frac{G(1-z)}{G(1+z)} \right)= 2\pi z\log\left(\frac{\sin\pi z}{\pi} \right)+\text{Cl}_2(2\pi z)\, . \, \Box </math>
 
 
Using the relation <math>\, G(1+z)=\Gamma(z)\, G(z) \,</math> and dividing the reflection formula by a factor of <math>\, 2\pi \,</math> gives the equivalent form:
 
 
:<math> \log\left( \frac{G(1-z)}{G(z)} \right)= z\log\left(\frac{\sin\pi z}{\pi}
\right)+\log\Gamma(z)+\frac{1}{2\pi}\text{Cl}_2(2\pi z) </math>
 
 
Ref: see '''Adamchik''' below for an equivalent form of the [[reflection formula]], but with a different proof.
 
==Reflection formula 2.0==
 
 
Replacing '''z''' with '''(1/2)-z''' in the previous reflection formula gives,  after some simplification, the equivalent formula shown below (involving [[Bernoulli polynomials]]):
 
 
:<math>\log\left( \frac{ G\left(\frac{1}{2}+z\right) }{ G\left(\frac{1}{2}-z\right) } \right) =</math>
 
 
:<math>
\log \Gamma \left(\frac{1}{2}-z \right) + B_1(z) \log 2\pi-\frac{1}{2}\log 2+\pi \int_0^z B_1(x) \tan \pi x \,dx</math>
 
 
 
==Taylor series expansion==
 
 
By [[Taylor's theorem]], and considering the logarithmic [[derivatives]] of the Barnes function, the following series expansion can be obtained:
 
 
:<math>\log G(1+z)= \frac{z}{2}\log 2\pi -\left( \frac{z+(1+\gamma)z^2}{2} \right) + \sum_{k=2}^{\infty}(-1)^k\frac{\zeta(k)}{k+1}z^{k+1}</math>
 
 
It is valid for <math>\, 0 < z < 1 \,</math>. Here, <math>\, \zeta(x) \,</math> is the [[Riemann Zeta function]]:
 
 
:<math> \zeta(x)=\sum_{k=1}^{\infty}\frac{1}{k^x} </math>
 
 
Exponentiating both sides of the Taylor expansion gives:
 
 
:<math> G(1+z)=\exp \left[ \frac{z}{2}\log 2\pi -\left( \frac{z+(1+\gamma)z^2}{2} \right) + \sum_{k=2}^{\infty}(-1)^k\frac{\zeta(k)}{k+1}z^{k+1} \right]=</math>
 
 
:<math>(2\pi)^{z/2}\text{exp}\left( -\frac{z+(1+\gamma)z^2}{2} \right) \exp \left[\sum_{k=2}^{\infty}(-1)^k\frac{\zeta(k)}{k+1}z^{k+1} \right]</math>
 
 
Comparing this with the [[Weierstrass product]] form of the Barnes function gives the following relation:
 
 
:<math>\exp \left[\sum_{k=2}^{\infty}(-1)^k\frac{\zeta(k)}{k+1}z^{k+1} \right] = \prod_{k=1}^{\infty}\left(1+\frac{z}{k}\right)^k\text{exp}\left(\frac{z^2}{2k}-z\right)</math>
 
==Multiplication formula==
 
 
Like the Gamma function, the G-function also has a multiplication formula:<ref>I. Vardi, ''Determinants of Laplacians and multiple gamma functions'', SIAM J. Math. Anal. '''19''', 493–507 (1988).</ref>
 
 
:<math>
G(nz)= K(n) n^{n^{2}z^{2}/2-nz} (2\pi)^{-\frac{n^2-n}{2}z}\prod_{i=0}^{n-1}\prod_{j=0}^{n-1}G\left(z+\frac{i+j}{n}\right)
</math>
 
 
where <math>K(n)</math> is a constant given by:
 
 
:<math> K(n)= e^{-(n^2-1)\zeta^\prime(-1)} \cdot
n^{\frac{5}{12}}\cdot(2\pi)^{(n-1)/2}\,=\,
(Ae^{-\frac{1}{12}})^{n^2-1}\cdot n^{\frac{5}{12}}\cdot (2\pi)^{(n-1)/2}.</math>
 
 
Here <math>\zeta^\prime</math> is the derivative of the [[Riemann zeta function]] and <math>A</math> is the [[Glaisher–Kinkelin constant]].
 
== Asymptotic expansion ==
 
 
The [[logarithm]] of ''G''(''z'' + 1) has the following asymptotic expansion, as established by Barnes:
 
 
:<math> \log G(z+1)=</math>
 
 
:<math>\frac{1}{12}~-~\log A~+~\frac{z}{2}\log 2\pi~+~\left(\frac{z^2}{2} -\frac{1}{12}\right)\log z~-~\frac{3z^2}{4}~+~
\sum_{k=1}^{N}\frac{B_{2k + 2}}{4k\left(k + 1\right)z^{2k}}~+~O\left(\frac{1}{z^{2N + 2}}\right).</math>
 
 
Here the <math>B_{k}</math> are the [[Bernoulli numbers]] and <math>A</math> is the [[Glaisher–Kinkelin constant]]. (Note that somewhat confusingly at the time of Barnes <ref>E.T.Whittaker and G.N.Watson, "A course of modern analysis", CUP.</ref> the [[Bernoulli number]] <math>B_{2k}</math> would have been written as <math>(-1)^{k+1} B_k </math>, but this convention is no longer current.) This expansion is valid for <math>z </math> in any sector not containing the negative real axis with <math>|z|</math> large.
 
==Relation to the Loggamma integral==
 
 
The parametric Loggamma can be evaluated in terms of the Barnes G-function (Ref: this result is found in '''Adamchik''' below, but stated without proof):
 
 
:<math> \int_0^z \log \Gamma(x)\,dx=\frac{z(1-z)}{2}+\frac{z}{2}\log 2\pi +z\log\Gamma(z) -\log G(1+z) </math>
 
 
The proof is somewhat indirect, and involves first considering the logarithmic difference of the [[Gamma function]] and Barnes G-function:
 
:<math>z\log \Gamma(z)-\log G(1+z)</math>
 
 
Where
 
 
:<math>\frac{1}{\Gamma(z)}= z e^{\gamma} \prod_{k=1}^{\infty}\left(1+\frac{z}{k}\right)e^{-z/k}</math>
 
 
and <math>\,\gamma\,</math> is the [[Euler-Mascheroni constant]].
 
 
Taking the logarithm of the [[Weierstrass product]] forms of the Barnes function and Gamma function gives:
 
 
:<math>z\log \Gamma(z)-\log G(1+z)=-z \log\left(\frac{1}{\Gamma (z)}\right)-\log G(1+z)=</math>
 
 
:<math>-z \left[ \log z+\gamma z +\sum_{k=1}^{\infty} \Bigg\{ \log\left(1+\frac{z}{k} \right) -\frac{z}{k} \Bigg\} \right]</math>
 
 
:<math>-\left[ \frac{z}{2}\log 2\pi -\frac{z}{2}-\frac{z^2}{2} -\frac{z^2 \gamma}{2} + \sum_{k=1}^{\infty} \Bigg\{k\log\left(1+\frac{z}{k}\right) +\frac{z^2}{2k} -z \Bigg\} \right]</math>
 
 
A little simplification and re-ordering of terms gives the series expansion:
 
 
:<math> \sum_{k=1}^{\infty} \Bigg\{ (k+z)\log \left(1+\frac{z}{k}\right)-\frac{z^2}{2k}-z \Bigg\}=</math>
 
 
:<math>-z\log z-\frac{z}{2}\log 2\pi +\frac{z}{2} +\frac{z^2}{2}- \frac{z^2 \gamma}{2}- z\log\Gamma(z) +\log G(1+z)</math>
 
 
Finally, take the logarithm of the [[Weierstrass product]] form of the [[Gamma function]], and integrate over the interval <math>\, [0,\,z]\, </math> to obtain:
 
 
:<math>\int_0^z\log\Gamma(x)\,dx=-\int_0^z \log\left(\frac{1}{\Gamma(x)}\right)\,dx=</math>
 
 
:<math>-(z\log z-z)-\frac{z^2 \gamma}{2}- \sum_{k=1}^{\infty} \Bigg\{ (k+z)\log \left(1+\frac{z}{k}\right)-\frac{z^2}{2k}-z \Bigg\}</math>
 
 
Equating the two evaluations completes the proof:
 
 
:<math> \int_0^z \log \Gamma(x)\,dx=\frac{z(1-z)}{2}+\frac{z}{2}\log 2\pi +z\log\Gamma(z) -\log G(1+z)\, . \, \Box</math>
 
==References==
<references/>
 
*{{dlmf|first=R.A. |last=Askey|first2=R.|last2=Roy|id=5.17}}
{{DEFAULTSORT:Barnes G-Function}}
[[Category:Number theory]]
[[Category:Special functions]]
 
*{{cite web|last=Adamchik|first=Viktor S.|title=Contributions to the Theory of the Barnes function|url=http://arxiv.org/pdf/math/0308086v1.pdf|accessdate=2003}}

Revision as of 22:01, 2 November 2013

In mathematics, the Barnes G-function G(z) is a function that is an extension of superfactorials to the complex numbers. It is related to the Gamma function, the K-function and the Glaisher–Kinkelin constant, and was named after mathematician Ernest William Barnes.[1] Up to elementary factors, it is a special case of the double gamma function.


Formally, the Barnes G-function is defined in the following Weierstrass product form:


G(1+z)=(2π)z/2exp(−z+z2(1+γ)2)∏k=1∞(1+zk)kexp(z22k−z)


where γ is the Euler–Mascheroni constant, exp(x) = ex, and ∏ is capital pi notation.


Functional equation and integer arguments

The Barnes G-function satisfies the functional equation


G(z+1)=Γ(z)G(z)


with normalisation G(1) = 1. Note the similarity between the functional equation of the Barnes G-function and that of the Euler Gamma function:


Γ(z+1)=zΓ(z)


The functional equation implies that G takes the following values at integer arguments:


G(n)={0if n=0,−1,−2,…∏i=0n−2i!if n=1,2,…


and thus

G(n)=(Γ(n))n−1K(n)


where Γ(x) denotes the Gamma function and K denotes the K-function. The functional equation uniquely defines the G function if the convexity condition: d3dx3G(x)≥0 is added.[2]

Reflection formula 1.0

The difference equation for the G function, in conjunction with the functional equation for the Gamma function, can be used to obtain the following reflection formula for the Barnes G function (originally proved by Hermann Kinkelin):


log⁡G(1−z)=log⁡G(1+z)−zlog⁡2π+∫0zπxcot⁡πxdx.


The logtangent integral on the right-hand side can be evaluated in terms of the Clausen function (of order 2), as is shown below:


2πlog⁡(G(1−z)G(1+z))=2πzlog⁡(sin⁡πzπ)+Cl2(2πz)


The proof of this result hinges on the following evaluation of the cotangent integral: introducing the notation Lc(z) for the logtangent integral, and using the fact that (d/dx)log⁡(sin⁡πx)=πcot⁡πx, an integration by parts gives


Lc(z)=∫0zπxlog⁡(sin⁡πx)dx=zlog⁡(sin⁡πz)−∫0zlog⁡(sin⁡πx)dx=


zlog⁡(sin⁡πz)−∫0z[log⁡(2sin⁡πx)−log⁡2]dx=


zlog⁡(2sin⁡πz)−∫0zlog⁡(2sin⁡πx)dx


Performing the integral substitution y=2πx⇒dx=dy/(2π) gives


zlog⁡(2sin⁡πz)−12π∫02πzlog⁡(2sin⁡πy2)dy


The Clausen function - of second order - has the integral representation


Cl2(θ)=−∫0θlog⁡|2sin⁡x2|dx


However, within the interval 0<θ<2π, the absolute value sign within the integrand can be omitted, since within the range the 'half-sine' function in the integral is strictly positive, and strictly non-zero. Comparing this definition with the result above for the logtangent itegral, the following relation clearly holds:


Lc(z)=zlog⁡(2sin⁡πz)+12πCl2(2πz)


Thus, after a slight rearrangement of terms, the proof is complete:


2πlog⁡(G(1−z)G(1+z))=2πzlog⁡(sin⁡πzπ)+Cl2(2πz).◻


Using the relation G(1+z)=Γ(z)G(z) and dividing the reflection formula by a factor of 2π gives the equivalent form:


log⁡(G(1−z)G(z))=zlog⁡(sin⁡πzπ)+log⁡Γ(z)+12πCl2(2πz)


Ref: see Adamchik below for an equivalent form of the reflection formula, but with a different proof.

Reflection formula 2.0

Replacing z with (1/2)-z in the previous reflection formula gives, after some simplification, the equivalent formula shown below (involving Bernoulli polynomials):


log⁡(G(12+z)G(12−z))=


log⁡Γ(12−z)+B1(z)log⁡2π−12log⁡2+π∫0zB1(x)tan⁡πxdx


Taylor series expansion

By Taylor's theorem, and considering the logarithmic derivatives of the Barnes function, the following series expansion can be obtained:


log⁡G(1+z)=z2log⁡2π−(z+(1+γ)z22)+∑k=2∞(−1)kζ(k)k+1zk+1


It is valid for 0<z<1. Here, ζ(x) is the Riemann Zeta function:


ζ(x)=∑k=1∞1kx


Exponentiating both sides of the Taylor expansion gives:


G(1+z)=exp⁡[z2log⁡2π−(z+(1+γ)z22)+∑k=2∞(−1)kζ(k)k+1zk+1]=


(2π)z/2exp(−z+(1+γ)z22)exp⁡[∑k=2∞(−1)kζ(k)k+1zk+1]


Comparing this with the Weierstrass product form of the Barnes function gives the following relation:


exp⁡[∑k=2∞(−1)kζ(k)k+1zk+1]=∏k=1∞(1+zk)kexp(z22k−z)

Multiplication formula

Like the Gamma function, the G-function also has a multiplication formula:[3]


G(nz)=K(n)nn2z2/2−nz(2π)−n2−n2z∏i=0n−1∏j=0n−1G(z+i+jn)


where K(n) is a constant given by:


K(n)=e−(n2−1)ζ′(−1)⋅n512⋅(2π)(n−1)/2=(Ae−112)n2−1⋅n512⋅(2π)(n−1)/2.


Here ζ′ is the derivative of the Riemann zeta function and A is the Glaisher–Kinkelin constant.

Asymptotic expansion

The logarithm of G(z + 1) has the following asymptotic expansion, as established by Barnes:


log⁡G(z+1)=


112−log⁡A+z2log⁡2π+(z22−112)log⁡z−3z24+∑k=1NB2k+24k(k+1)z2k+O(1z2N+2).


Here the Bk are the Bernoulli numbers and A is the Glaisher–Kinkelin constant. (Note that somewhat confusingly at the time of Barnes [4] the Bernoulli number B2k would have been written as (−1)k+1Bk, but this convention is no longer current.) This expansion is valid for z in any sector not containing the negative real axis with |z| large.

Relation to the Loggamma integral

The parametric Loggamma can be evaluated in terms of the Barnes G-function (Ref: this result is found in Adamchik below, but stated without proof):


∫0zlog⁡Γ(x)dx=z(1−z)2+z2log⁡2π+zlog⁡Γ(z)−log⁡G(1+z)


The proof is somewhat indirect, and involves first considering the logarithmic difference of the Gamma function and Barnes G-function:

zlog⁡Γ(z)−log⁡G(1+z)


Where


1Γ(z)=zeγ∏k=1∞(1+zk)e−z/k


and γ is the Euler-Mascheroni constant.


Taking the logarithm of the Weierstrass product forms of the Barnes function and Gamma function gives:


zlog⁡Γ(z)−log⁡G(1+z)=−zlog⁡(1Γ(z))−log⁡G(1+z)=


−z[log⁡z+γz+∑k=1∞{log⁡(1+zk)−zk}]


−[z2log⁡2π−z2−z22−z2γ2+∑k=1∞{klog⁡(1+zk)+z22k−z}]


A little simplification and re-ordering of terms gives the series expansion:


∑k=1∞{(k+z)log⁡(1+zk)−z22k−z}=


−zlog⁡z−z2log⁡2π+z2+z22−z2γ2−zlog⁡Γ(z)+log⁡G(1+z)


Finally, take the logarithm of the Weierstrass product form of the Gamma function, and integrate over the interval [0,z] to obtain:


∫0zlog⁡Γ(x)dx=−∫0zlog⁡(1Γ(x))dx=


−(zlog⁡z−z)−z2γ2−∑k=1∞{(k+z)log⁡(1+zk)−z22k−z}


Equating the two evaluations completes the proof:


∫0zlog⁡Γ(x)dx=z(1−z)2+z2log⁡2π+zlog⁡Γ(z)−log⁡G(1+z).◻

References

  1. ↑ E.W.Barnes, "The theory of the G-function", Quarterly Journ. Pure and Appl. Math. 31 (1900), 264–314.
  2. ↑ M. F. Vignéras, L'équation fonctionelle de la fonction zêta de Selberg du groupe mudulaire SL(2,ℤ), Astérisque 61, 235–249 (1979).
  3. ↑ I. Vardi, Determinants of Laplacians and multiple gamma functions, SIAM J. Math. Anal. 19, 493–507 (1988).
  4. ↑ E.T.Whittaker and G.N.Watson, "A course of modern analysis", CUP.