## Abstract

The method of brackets is a method of integration based upon a small number of heuristic rules. Some of these have been made rigorous. An example of an integral involving the Bessel function is used to motivate a new evaluation rule.

Show Summary Details# Pochhammer symbol with negative indices. A new rule for the method of brackets

#### Open Access

## Abstract

## 1 Introduction

## 2 The method of brackets

## 3 A first evaluation of entry 6:671:7 in Gradshteyn and Ryzhik

## 4 An alternative evaluation

## 5 Extensions of the Pochhammer symbol

## Acknowledgement

## References

## About the article

In This Section# Open Mathematics

### formerly Central European Journal of Mathematics

In This Section

Editor-in-Chief: Gianazza, Ugo / Vespri, Vincenzo

1 Issue per year

IMPACT FACTOR 2016 (Open Mathematics): 0.682

IMPACT FACTOR 2016 (Central European Journal of Mathematics): 0.489

CiteScore 2016: 0.62

SCImago Journal Rank (SJR) 2015: 0.521

Source Normalized Impact per Paper (SNIP) 2015: 1.233

Mathematical Citation Quotient (MCQ) 2015: 0.39

The method of brackets is a method of integration based upon a small number of heuristic rules. Some of these have been made rigorous. An example of an integral involving the Bessel function is used to motivate a new evaluation rule.

Key Words: Method of brackets; Pochhammer symbol; Bessel integral

The evaluation of definite integrals is connected, in a surprising manner, to many topics in Mathematics. The last author has described in [15] and [17] some of these connections. Many of these evaluations appear in the classical table [13] and proofs of these entries have appeared in a series of papers starting with [16] and the latest one is [1]. An interesting new method of integration, developed in [10] in the context of integrals coming from Feynman diagrams, is illustrated here in the evaluation of and entry from the classical table mentioned above. Each of these entries have an assigned number and the one discussed here is 6:671:7 in [13]: $$I\text{\hspace{0.17em}}:=\underset{0}{\overset{\infty}{{\displaystyle \int}}}{J}_{0}(ax)sin(bx)\text{\hspace{0.17em}}dx\text{\hspace{0.17em}}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\text{\hspace{0.17em}}\{\begin{array}{ll}0\hfill & \text{if}\text{\hspace{0.17em}}0<b<a,\hfill \\ 1/\sqrt{{b}^{2}-{a}^{2}},\hfill & \text{if}\text{\hspace{0.17em}}0<a<b.\hfill \end{array}$$(1)

Here *J*_{0} is the Bessel function. The reader is referred to [3] for an introduction to these classical functions.

The new integration procedure has been named *the method of brackets*. It is based upon a small number of heuristic rules. This method is close in spirit to the classical *umbral method*, developed in the Combinatorial world by G. C. Rota. The reader will find in [18] an elementary introduction to this topic. The application of this method to Special Functions may be found, for instance, in a series of papers by G. Dattoli and collaborators, such as [4], [5] and [6]. The connections between the umbral method and the method of brackets is the subject of current work and results will be reported elsewhere.

The rules for the method of brackets are described in the next section. Sections 3 and 4 present the evaluation of (1). It turns out that a *naive limiting process* leads to a conflict between these two evaluations. The issue comes from a limit computation involving the Pochhammer symbol (*a*)_{k}. It is the goal of this paper to clarify this conflict. This is presented in Section 5 leading to the proposal of a new rule dealing with the extension of the Pochhammer symbol to negative integer indices.

A method to evaluate integrals over the half line [0, ∞), based on a small number of rules has been developed in [10–12]. This method of brackets is described next. The heuristic rules are currently being made rigorous in [2] and [14]. The reader will find in [7–9] a large collection of evaluations of definite integrals that illustrate the power and flexibility of this method.

For *a* ∈ ℂ, the symbol
$$\u3008a\u3009\text{\hspace{0.17em}}\mapsto \text{\hspace{0.17em}}\underset{0}{\overset{\infty}{{\displaystyle \int}}}{x}^{a-1}\text{\hspace{0.17em}}dx$$(2)
is the *bracket* associated to the (divergent) integral on the right. The symbol
$${\varphi}_{n}\text{\hspace{0.17em}}\text{\hspace{0.17em}}:=\frac{{(-1)}^{n}}{\text{\Gamma}(n+1)}\text{\hspace{0.17em}}$$(3)
is called the *indicator* associated to the index *n*. The notation *ϕ*_{i1i2...ir}, or simply *ϕ*_{12...r}, denotes the product *ϕ*_{i1} *ϕ*_{i2} ... *ϕ*_{ir}.

**Rules for the production of bracket series**

**Rule P _{1}**. If the function

**Rule P _{2}**. For

**Rules for the evaluation of a bracket series**

**Rule E _{1}** The one-dimensional bracket series is assigned the value
$$\sum}_{n}{\varphi}_{n}\text{\hspace{0.17em}}f(n)\u3008an\text{\hspace{0.17em}}+\text{\hspace{0.17em}}b\u3009\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\frac{1}{\left|a\right|}f({n}^{\ast})\text{\Gamma}(-{n}^{\ast}),$$(7)
where

The next rule provides a value for multi-dimensional bracket series where the number of sums is equal to the number of brackets.

**Rule E _{2}** Assume the matrix

**Rule E _{3}** Each representation of an integral by a bracket series has associated an

index = number of sums – number of brackets. (8)

It is important to observe that the index is attached to a specific representation of the integral and not just to integral itself. The experience obtained by the authors using this method suggests that, among all representations of an integral as a bracket series, the one with *minimal index* should be chosen.

The value of a multi-dimensional bracket series of positive index is obtained by computing all the contributions of maximal rank by Rule *E*_{2}. These contributions to the integral appear as series in the free parameters. Series converging in a common region are added and divergent series are discarded.

**Note 2.1***The rule* *E*_{1}*is a version of the Ramanujan’s Master Theorem. A complete justification of this rule is provided in [2]. The justification of the remaining rules is the subject of current work*.

**Note 2.2***A systematic procedure in the simplification of the series obtained by this procedure has been used throughout the literature: express factorials in terms of the gamma function and the transform quotients of gamma terms into Pochhammer symbol, defined by*
$${(a)}_{k}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\frac{\text{\Gamma}(a+k)}{\text{\Gamma}(a)}.$$(9)

*Any presence of a Pochhammer with a negative index k is transformed by the rule*
$${(a)}_{-k}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\frac{{(-1)}^{k}}{{(1-a)}_{k}}.$$(10)

The example discussed in the next two section provides motivation for an additional evaluation rule for the method of brackets.

The evaluation of (1) uses the series $${J}_{0}(ax)\text{\hspace{0.17em}}=\text{\hspace{0.17em}}{\displaystyle \sum}_{m=0}^{\infty}{\varphi}_{m}\frac{{a}^{2m}}{\text{\Gamma}(m+1){2}^{2m}}{x}^{2m}$$(11) and $$\text{sin}(bx)\text{\hspace{0.17em}}=\text{\hspace{0.17em}}{\displaystyle \sum}_{n=0}^{\infty}{\varphi}_{n}\frac{\text{\Gamma}(n+1)}{\text{\Gamma}(2n+2)}{b}^{2n+1}{x}^{2n+1}.$$(12)

Therefore the integral in (1) is given by $$I\text{\hspace{0.17em}}=\text{\hspace{0.17em}}{\displaystyle \sum}_{m\text{\hspace{0.17em}},n}{\varphi}_{m\text{\hspace{0.17em}},n}\frac{{a}^{2m}{b}^{2n+1}\text{\Gamma}(n\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)}{{2}^{2m}\text{\Gamma}(m\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)\text{\Gamma}(2n\text{\hspace{0.17em}}+\text{\hspace{0.17em}}2)}\u30082m\text{\hspace{0.17em}}+\text{\hspace{0.17em}}2n\text{\hspace{0.17em}}+\text{\hspace{0.17em}}2\u3009.$$(13)

The duplication formula for the gamma function transforms this expression to $$I\text{\hspace{0.17em}}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\text{\hspace{0.17em}}\frac{\sqrt{\pi}}{2}\text{\hspace{0.17em}}{\displaystyle \sum}_{m\text{\hspace{0.17em}},n}{\varphi}_{m\text{\hspace{0.17em}},n}\frac{{a}^{2m}{b}^{2n+1}}{{2}^{2m+2n}\text{\Gamma}(m\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)\text{\Gamma}(n\text{\hspace{0.17em}}+\text{\hspace{0.17em}}3/2)}\u30082m\text{\hspace{0.17em}}+\text{\hspace{0.17em}}2n\text{\hspace{0.17em}}+\text{\hspace{0.17em}}2\u3009.$$(14)

Eliminating the parameter *n* using Rule *E*_{1} gives *n*^{*} = –*m* – 1 and produces
$$\begin{array}{lll}I\hfill & =\hfill & \frac{\pi}{2}\text{\hspace{0.17em}}{\displaystyle \sum _{m=0}^{\infty}\text{\hspace{0.17em}}{\varphi}_{m}\frac{1}{\text{\Gamma}\left(-m\text{\hspace{0.17em}}+\text{\hspace{0.17em}}\frac{1}{2}\right)}}{\left(\frac{a}{b}\right)}^{2m}\hfill \\ \hfill & =\hfill & \frac{1}{b}\text{\hspace{0.17em}}{\displaystyle \sum _{m=0}^{\infty}\text{\hspace{0.17em}}\frac{{(-1)}^{m}}{m!}\frac{{\left(\frac{{a}^{2}}{{b}^{2}}\right)}^{m}}{{\left(\frac{1}{2}\right)}_{-m}}}\hfill \\ \hfill & =\hfill & \frac{1}{b}\text{\hspace{0.17em}}{\displaystyle \sum _{m=0}^{\infty}\text{\hspace{0.17em}}{\left(\frac{1}{2}\right)}_{m}\text{\hspace{0.17em}}\frac{1}{m!}\text{\hspace{0.17em}}{\left(\frac{{a}^{2}}{{b}^{2}}\right)}^{m}}\hfill \\ \hfill & =\hfill & \frac{1}{\sqrt{{b}^{2}-{a}^{2}}}.\hfill \end{array}$$

The condition |*b*| > |*a*| is imposed to guarantee the convergence of the series on the third line of the previous argument.

The series obtained by eliminating the parameter *m* by *m*^{*} = –*n* – 1 vanishes because of the factor Γ(*m* + 1) in the denominator. The formula has been established.

A second evaluation of (1) begins with $$\underset{0}{\overset{\infty}{{\displaystyle \int}}}\text{\hspace{0.17em}}{J}_{0}(ax)\text{\hspace{0.17em}}\mathrm{sin}(bx)\text{\hspace{0.17em}}dx\text{\hspace{0.17em}}=\text{\hspace{0.17em}}{\displaystyle \sum}_{m\text{\hspace{0.17em}},n}{\varphi}_{m\text{\hspace{0.17em}},n}\frac{{a}^{2m}{b}^{2n+1}}{{2}^{2m}m!(2n\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)!}\text{\hspace{0.17em}}\u30082m\text{\hspace{0.17em}}+\text{\hspace{0.17em}}2n\text{\hspace{0.17em}}+\text{\hspace{0.17em}}2\u3009.$$(15)

The evaluation of the bracket series is described next.

*Case 1*. Choose *n* as the free parameter. Then *m*^{*} = –*n* – 1 and the contribution to the integral is
$${I}_{1\text{\hspace{0.17em}}}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\frac{1}{2}\text{\hspace{0.17em}}{\displaystyle \sum}_{n=0}^{\infty}\text{\hspace{0.17em}}{\varphi}_{n}{\left(\frac{a}{2}\right)}^{-2n-2}{b}^{2n+1}\frac{1}{\text{\Gamma}(-n)}\frac{n!}{(2n\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)!}\text{\Gamma}(n\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1).$$(16)

Each term in the sum vanishes because the gamma function has a pole at the negative integers.

*Case 2*. Choose *m* as a free parameter. Then *2m* + *2n* + 2 = 0 gives *n*^{*} = –*m* – 1. The contribution to the integral is
$${I}_{2\text{\hspace{0.17em}}}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\frac{1}{2}\text{\hspace{0.17em}}{\displaystyle \sum}_{m}{\varphi}_{m}{\left(\frac{a}{2}\right)}^{2m}{b}^{-2m\text{\hspace{0.17em}}-1}\frac{1}{m!}\cdot \frac{n!}{(2n\text{\hspace{0.17em}}\text{\hspace{0.17em}}+\text{\hspace{0.17em}}\text{\hspace{0.17em}}1)!}{|}_{n\text{\hspace{0.17em}}=\text{\hspace{0.17em}}-m\text{\hspace{0.17em}}-1}\text{\Gamma}(m\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1).$$(17)

Now write $$\frac{n!}{(2n\text{\hspace{0.17em}}\text{\hspace{0.17em}}+\text{\hspace{0.17em}}\text{\hspace{0.17em}}1)!}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\frac{1}{{(n\text{\hspace{0.17em}}\text{\hspace{0.17em}}+\text{\hspace{0.17em}}\text{\hspace{0.17em}}1)}_{n\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1}}$$(18) and (17) becomes $${I}_{2\text{\hspace{0.17em}}}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\text{\hspace{0.17em}}\frac{1}{2b}\text{\hspace{0.17em}}\text{\hspace{0.17em}}{\displaystyle \sum}_{m=0}^{\infty}\text{\hspace{0.17em}}\frac{{(-1)}^{m}}{m!}\text{\hspace{0.17em}}{\left(\frac{a}{2b}\right)}^{2m}\frac{1}{{(-m)}_{-m}}.$$(19)

Transforming the term (*–m) _{–m}* by using
$${(x)}_{-n}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\frac{{(-1)}^{n}}{{(1-x)}_{n}}.$$(20)
gives
$${(-m)}_{-m}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\frac{{(-1)}^{m}}{{(1\text{\hspace{0.17em}}+\text{\hspace{0.17em}}m)}_{m}}.$$(21)

Replacing in (19) produces $$\begin{array}{lll}{I}_{2}\hfill & =\hfill & \frac{1}{2b}\text{\hspace{0.17em}}{\displaystyle \sum}_{k=0}^{\infty}\text{\hspace{0.17em}}{\left(\frac{1}{2}\right)}_{k}\frac{1}{k!}\text{\hspace{0.17em}}{\left(\frac{{a}^{2}}{{b}^{2}}\right)}^{k}\hfill \\ \hfill & =\hfill & \frac{1}{2}\frac{1}{\sqrt{{b}^{2}\text{\hspace{0.17em}}-\text{\hspace{0.17em}}{a}^{2}}}.\hfill \end{array}$$

The method of brackets produces *half of the expected answer*. Naturally, it is possible that the entry in [13] is erroneous (this happens once in a while). Some numerical computations and the evaluation in the previous section, should convince the reader that this is not the case. The source of the error is the use of (20) for the evaluation of the term (*–m) _{–m}*. A discussion is presented in Section 5.

Rule *E*_{1} of the method of brackets requires the evaluation of *f(n ^{*})*. In many instances, this involves the evaluation of the Pochhammer symbol (

The first extension of (*x*)_{m} to negative values of *n* comes from the identity
$${(x)}_{-m}\text{\hspace{0.17em}}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\text{\hspace{0.17em}}\frac{{(-1)}^{m}}{{(1-x)}^{m}}.$$(23)

This is obtained from
$${(x)}_{-m}\text{\hspace{0.17em}}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\text{\hspace{0.17em}}\frac{\text{\Gamma}(x-m)}{\text{\Gamma}(x)}\text{\hspace{0.17em}}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\text{\hspace{0.17em}}\frac{\text{\Gamma}(x-m)}{(x\text{\hspace{0.17em}}-\text{\hspace{0.17em}}1)(x\text{\hspace{0.17em}}-\text{\hspace{0.17em}}2)\cdots (x\text{\hspace{0.17em}}-\text{\hspace{0.17em}}m)\text{\Gamma}(x-m)}$$(24)
and then changing the signs of each of the factors. This is valid as long as *x* is not a negative integer. The limiting value of the right-hand side in (23) as *x* → –*km*, with *k* ∈ ℕ, is
$${(k\text{\hspace{0.17em}}m)}_{-m}\text{\hspace{0.17em}}\text{\hspace{0.17em}}=\text{\hspace{0.17em}}\text{\hspace{0.17em}}\frac{{(-1)}^{m}(k\text{\hspace{0.17em}}m)!}{((k\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)m)!}.$$(25)

On the other hand, the limiting value of the left-hand side is $$\begin{array}{lll}\underset{\epsilon \to 0}{\mathrm{lim}}\text{\hspace{0.17em}}{(-k(m\text{\hspace{0.17em}}+\text{\hspace{0.17em}}\epsilon ))}_{-(m\text{\hspace{0.17em}}+\text{\hspace{0.17em}}\epsilon )}\hfill & =\hfill & \underset{\epsilon \to 0}{\mathrm{lim}}\text{\hspace{0.17em}}\frac{\text{\Gamma}(-(k\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)m\text{\hspace{0.17em}}-\text{\hspace{0.17em}}(k\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)\epsilon )}{\text{\Gamma}(-km\text{\hspace{0.17em}}+\text{\hspace{0.17em}}k\epsilon )}\hfill \\ \hfill & =\hfill & \underset{\epsilon \to 0}{\mathrm{lim}}\text{\hspace{0.17em}}\frac{\text{\Gamma}(-(k\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)\epsilon ){(-\text{\hspace{0.17em}}(k\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)\epsilon )}_{-(k\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)m}}{\text{\Gamma}(-k\epsilon ){(-k\epsilon )}_{-km}\text{\hspace{0.17em}}}\hfill \\ \hfill & =\hfill & \underset{\epsilon \to 0}{\mathrm{lim}}\text{\hspace{0.17em}}\frac{\text{\Gamma}(-(k\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)\epsilon )}{\text{\Gamma}(-k\epsilon )}\frac{{(-1)}^{(k\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)m}}{{(1\text{\hspace{0.17em}}+\text{\hspace{0.17em}}(k\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)\epsilon )}_{(k\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)m}}\cdot \frac{{(1\text{\hspace{0.17em}}+\text{\hspace{0.17em}}k\epsilon )}_{km}}{{(-1)}^{km}}\hfill \\ \hfill & =\hfill & \frac{{(-1)}^{m}(km)!}{((k\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1)m)!}\cdot \text{\hspace{0.17em}}\frac{k}{k\text{}\text{\hspace{0.17em}}+\text{\hspace{0.17em}}1}.\hfill \end{array}$$

Therefore the function (*x) _{–m}* is discontinuous at

For *k* = 1, this ratio becomes 2. This explains the missing ½ in the calculation in Section 4. Therefore it is the discontinuity of (10) at negative integer values of the variables, what is responsible for the error in the evaluation of the integral (1).

This example suggest that the rules of the method of brackets should be supplemented with an additional one:

**Rule E _{4}** Let

A variety of other examples confirm that this heuristic rule leads to correct evaluations.

The last author acknowledges the partial support of NSF-DMS 1112656. The second author is a graduate student, partially supported by the same grant.

- [1]
Amdeberhan T., Dixit A., Guan X., Jiu L., Kuznetsov A., Moll V., Vignat C., The integrals in Gradshteyn and Ryzhik. Part 30: Trigonometric integrals. Scientia, 27:47-74, 2016

- [2]
Amdeberhan T., Espinosa O., Gonzalez, Harrison M., Moll V., Straub A., Ramanujan Master Theorem. The Ramanujan Journal, 29:103-120, 2012

- [3]
Andrews G.E., Askey R., Roy R., Special Functions, volume 71 of Encyclopedia of Mathematics and its Applications. Cambridge University Press, New York, 1999

- [4]
Dattoli G., Cesarano C., Saccetti D., Miscellaneous results on the generating functions of special functions. Integral Transforms and Special Functions, 12:315-322, 2001

- [5]
Dattoli G., Ricci P., Cesarano C., The Lagrange polynomials, the associated generalizations, and the Umbral Calculus. Integral Transforms and Special Functions, 14:181-186, 2003

- [6]
Dattoli G., Ricci P., Cesarano C., Beyond the monomiality: the Monumbrality Principle. Jour. Comp. Anal. Appl., 6:77-83, 2004

- [7]
Gonzalez I., Kohl K., Moll V., Evaluation of entries in Gradshteyn and Ryzhik employing the method of brackets. Scientia, 25:65-84, 2014

- [8]
Gonzalez I., Moll V., Definite integrals by the method of brackets. Part 1. Adv. Appl. Math., 45:50-73, 2010

- [9]
Gonzalez I., Moll V., Straub A., The method of brackets. Part 2: Examples and applications. In T. Amdeberhan, L. Medina, and Victor H. Moll (Eds.), Gems in Experimental Mathematics, volume 517 of Contemporary Mathematics, pages 157-172. American Mathematical Society, 2010

- [10]
Gonzalez I., Schmidt I., Optimized negative dimensional integration method (NDIM) and multiloop Feynman diagram calculation. Nuclear Physics B, 769:124-173, 2007

- [11]
Gonzalez I., Schmidt I., Modular application of an integration by fractional expansion (IBFE) method to multiloop Feynman diagrams. Phys. Rev. D, 78:086003, 2008

- [12]
Gonzalez I., Schmidt I., Modular application of an integration by fractional expansion (IBFE) method to multiloop Feynman diagrams II. Phys. Rev. D, 79:126014, 2009

- [13]
Gradshteyn I.S., Ryzhik I.M., Table of Integrals, Series, and Products. Edited by D. Zwillinger and V. Moll. Academic Press, New York, 8th edition, 2015.POCHHAMMER SYMBOL WITH NEGATIVE INDICES 7

- [14]
Jiu L., On the method of brackets: rules, examples, interpretations and modifications. Submitted for publication, 2016

- [15]
Moll V., The evaluation of integrals: a personal story. Notices of the AMS, 49:311-317, 2002

- [16]
Moll V., The integrals in Gradshteyn and Ryzhik. Part 1: A family of logarithmic integrals. Scientia, 14:1-6, 2007

- [17]
Moll V., Seized opportunities. Notices of the AMS, pages 476-484, 2010

- [18]
Roman S., The Umbral Calculus. Dover, New York, 1984

**Received**: 2015-07-11

**Accepted**: 2016-09-08

**Published Online**: 2016-09-30

**Published in Print**: 2016-01-01

**Citation Information: **Open Mathematics, ISSN (Online) 2391-5455, DOI: https://doi.org/10.1515/math-2016-0063. Export Citation

## Comments (0)