Irish Math. Soc.

Bulletin
Number 86, Winter 2020, 109–113
ISSN 0791-5578

Cornel Ioan Vălean: (Almost) Impossible Integrals, Sums, and Series,

REVIEWED BY SEÁN M. STEWART

On first opening this book it did not take me long to realise there was something
Boschian about it. It is a vast store of results for definite integrals and infinite series
that at first glance seem to defy what is possible to achieve analytically. Those drawn
to the evaluation of integrals or series will find the present volume particularly difficult
to pass by. Evaluating integrals has always had its coterie of dedicated admirers. G. H.
Hardy once famously remarked ‘he could never resist the challenge of a definite integral’
[9, p. xi]. From the predilections of the current author it seems he would tend to have
to agree.
Forming part of Springer’s ‘Problem Books in Mathematics’ series the book is divided
into two parts. The first three chapters on integrals forms the first half of the book
(Chapter 1 the problems, Chapter 2 some hints, Chapter 3 their solutions), the last three
chapters on series and a few finite sums forms the second half of the book (Chapter 4 the
problems, Chapter 5 some hints, Chapter 6 their solutions). There are sixty problems
in each part. Classical but tricky integrals such as (Problem 1.7 (i))
log2 (1 + x)
Z 1
1
dx = ζ(3), (1)
0 x 4
where ζ denotes the Riemann zeta function make an appearance but there is much more
besides. As an example of the latter consider (Problem 1.33)
Z 1    
x x dx
Li2 Li2 .
0 x − 1 x + 1 x
Here Li2 is the dilogarithm. It is problems like this that live up to the title of the book
as being ‘almost’ impossible. Such an integral is sure to leave many scratching their
head and wondering how a closed-form solution is even possible. That it is I leave to
the interested reader.
Part of the fun and unusualness of the book is each of the 120 problems appear under
their own headings. Here one finds, for example (Problem 1.41)
A Little Integral-Beast from Inside Interesting Integrals Together with
a Similar Version of It Tamed by Real Methods
or (Problem 4.37)
Preparing the Weapons of The Master Theorem of Series to Breach
the Fortress of the Challenging Harmonic Series of Weight 7: The 1st
Episode
While the titles are delightfully curious and at times contain an interesting turn of
phase, a small quibble is in other places the text could have benefited from tighter
editing from the publisher given the first language of the author is not English.
Key words and phrases. Definite integrals, series, harmonic numbers, Euler sums, Riemann zeta
function, polylogarithms.
Received on 19-8-2020; revised 7-9-2020.

2020
c Irish Mathematical Society
109
110 SEÁN M. STEWART

As expected from a problem book there is little given in the way of preliminaries. It
is more a case of diving in head first and hoping you do not immediately hit the bottom.
A course in integral calculus is a given and one in the elements of real analysis would
be helpful. The reader is also expected to have a high degree of fluency with ‘standard’
special functions such as polylagarithms, the polygamma function, and the Riemann
zeta function, and to a lesser extent the Dirichlet beta function, the inverse tangent
integral, Lerch transcendent, and Legendre’s chi function of order two. Without knowl-
edge of these progress will be largely impossible. Some familiarity in the manipulation
of infinite series is also expected.
The first half of the book is devoted to definite integrals. Most of these are single
integrals with a handful of double integrals and the occasional triple integral. The
solutions given are comprehensive. An interesting feature is only real methods are
considered – there is no contour integration – and seems to be very much the preference
of the author. Sometimes two different methods are given and alternative approaches
are referenced. Occasionally interesting pieces of information about a particular integral
or its method of solution is included. As an example of this, for the first of the integrals
given in Problem 1.3 we learn it is believed to have first appeared in a book containing
a collection of problems suitable for the Cambridge course published in 1867 by the
English mathematician Joseph Wolstenholme (1829–1891) [12, p. 214]. Some of the
integrals are classical. Several were proposed by the author, having first appeared in
the problem sections of journals such as The American Mathematical Monthly and La
Gaceta de la Real Sociedad Matemática Española. Others still are completely new and
original.
The methods used in solving the integrals are varied. Some are well known, others
are inventive and creative. These include differentiating under the integral sign after the
introduction of a parameter; what is commonly known as ‘Feynman’s trick’; converting
a single integral to a double integral, and the use of infinite series. Still other ways is by
exploiting algebraic identities, employing symmetry, or by creating a system of relations
involving the integrals I and J and finding I + J and I − J first. As an example of the
first of these latter approaches, from the algebraic identity

(A + B)2 + (A − B)2 = 2A2 + 2B 2 ,

on setting A = log(1 − x) and B = log(1 + x) the integral in (1) can be found. Such an
ingenious approach was extensively used by De Doelder to evaluate many interesting
integrals [5]. On display is the author’s considerable manipulative dexterity and even
for those who may know how to solve a particular problem there is much one can learn
from a close reading of the provided solutions.
The second half of the book is devoted to the evaluation of series but it is really a
rumination by the author on a particular type of series known as Euler sums. These
are infinite series involving the harmonic numbers. As their name suggests, these series
are named after the great Swiss mathematician Leonhard Euler, who along with the
German mathematician Christian Goldbach initiated their study in the mid-eighteenth
century. The nth generalised harmonic number of order p ∈ N is defined by
n
X 1
Hn(p) = ,
kp
k=1

(p) (1)
such that H0 ≡ 0. When p = 1 we have Hn ≡ Hn , the nth harmonic number.
If we let π = (π1 , π2 , . . . , πk ) be a partition of integer p into k summands so that
p = π1 + · · · + πk , π1 6 π2 6 · · · 6 πk , and q is an integer such that q > 2, the classical
 Book Review 111

(non-linear) Euler sum is defined by [6, p. 16]

One of Euler’s famous early results is

Indeed, thirty-two different proofs of this classic result can be found in [4]. Later Euler
found a generalisation
∞ q−2
X Hn 1 1X
= (q + 2)ζ(q + 1) − ζ(n + 1)ζ(q − n),
nq 2 2
n=1 n=1

but it is disappointing to find this is not one of the problems given in the text (it
is however stated without proof; see Eq. (3.45) on page 87). Perhaps the author
considered the linear case far too elementary? A multitude of mostly non-linear Euler
sums or closely related series involving the harmonic numbers appear. Seven problems
towards the end deal specifically with alternating Euler sums.
My favourite of all the Euler sums is what is today referred to as the series of Au-
Yeung. It is (Problem 4.22)
∞ 
Hn 2 17 17π 4
X 
= ζ(4) = . (3)
n 4 360
n=1

Missed by Euler this series is largely responsible for the renewed fortunes of these sums.
Starting in the mid-1990s with the work of Bailey, Borwein, and Girgensohn [1] and
Borwein, Borwein, and Girgensohn [2] interest in series of this type was revived by an
accidental discovery of (3). On this latter point the Borwein brothers write [3, p. 1191]
This identity [namely (3)] was surprising and new to us when Enrico
Au-Yeung (an undergraduate student in the Faculty of Mathematics in
Waterloo) conjectured it on the basis of a computation of 500, 000 terms
(five digit accuracy!); our first impulse was to perform a higher-order
computation to show it to be false. It is not easy to naively compute
the value of the sum to more than about eight places.
The brothers went on to prove the result with the literature on Euler sums now vast.
Ironically the proof of (3) responsible for rejuvenating interest in Euler sums was itself
a rediscovery having first appeared, surprisingly, as a problem in the September 1948
issue of The American Mathematical Monthly [8].
From (2) we see a seemingly endless variety of Euler sums are possible. Many of
the problems found in the second half of the book are devoted to evaluating particular
examples of such sums. Beyond the literature little in the way of Euler sums has so
far found itself a place in modern texts. A handful of problems are given in [10, pp.
228–229] and a more substantial set, but at a level slightly easier than those found in
the current text, can be found in [7, pp. 148–151], which makes the present collection
a welcome addition. Having problems and their solutions for a large collection of Euler
sums in one location means the text should serve as a future source of reference for
sums of this type.
Not all the series problems appearing in the second half of the book are straight up
evaluation of Euler sums. More general and unusual sums containing a product between
112 SEÁN M. STEWART

the nth harmonic number and the tail of the Riemann zeta function can be found. One
example of this is (Problem 4.45 (i))
∞
π6
 
X Hn 1 1 5
ζ(4) − 1 − 4 − · · · − 4 = ζ(6) = .
n 2 n 48 9072
n=1

As was the case for the integrals presented in the first half of the book, solutions
to the series are found using only real methods. The methods used in their evaluation
are a combination of conversion to an integral (often one already found in the first
half of the book), the use of generating functions, or through series manipulation, of
which the author makes heavy use. Here Abel’s summation, changing the order of
double summations, and the author’s own Master Theorem of Series [11] are among
the various techniques used. In the hands of the present author these are a powerful
armamentaria.
I love Euler sums and would like to think they are useful. The author gives no hint
of their usefulness. They simply appear as distant peaks that need to be scaled and
conquered. In terms of usefulness they can often serve in the evaluation of integrals.
But must a use be found for everything? Who cannot help but marvel at a beautiful
closed-form solution to a problem many think is not possible? Euler sums may be a
minor but scenic tributary of modern day mathematics but is one deeply rooted in the
past that can be considered interesting for their own sake. If they are to eventually find
wide applicability the fact remains for this to be discovered by others.
For the most part the author’s manipulations with series and integrals are very clever
and a lot of fun to watch. At times an integral or series that seems impossible disappears
under a sea of what initially seems to be unrelated calculations only for it to re-emerge
several pages later with its victor clutching at it solution. At other times invoking a
hidden symmetry of the problem sees the initial impasse or but melt away. Many times
my reaction to this was one of wonderment. Where do people get such ideas from?
On encountering this we marvel that there are those who can imagine such things and
make such unexpected connections.
So what is one to make of the book? Some readers may feel they have been trans-
ported back in time, finding themselves negotiating some rich undiscovered vein of late
eighteenth or early nineteenth century mathematics. Pure mathematicians will proba-
bly grumble at the cavalier approach taken to rigour such as interchanges made between
infinite summations and integrations, but for the type of person this book is most likely
to appeal to this is but a small cavil. The book can serve as a useful supply of difficult
definite integrals and infinite series problems for undergraduates or as a useful starting
point for those wishing to attempt similar types of problems that arise from time to
time in the various journals with dedicated problem pages. And there are of course a
small group of those for whom the challenge of a definite integral is difficult to resist.
In this book they have found themselves the perfect antidote.

References
[1] D. H. Bailey, J. M. Borwein, R. Girgensohn: Experimental evaluation of Euler sums, Exp. Math.
(3) 1 (1994), 17–30.
[2] D. Borwein, J. M. Borwein, R. Girgensohn: Explicit evaluation of Euler sums, Proc. Edinb. Math.
Soc. (38) 2 (1995), 277–294.
[3] D. Borwein, J. M. Borwein: On an intriguing integral and some series related to ζ(4), Proc. Amer.
Math. Soc. (123) 4 (1995), 1191–1198.
[4] J. M. Borwein, D. M. Bradley: Thirty-two Goldbach variations, Int. J. Number Theory (2) 1 (2006),
65–103.
[5] P. J. De Doelder: On some series containing ψ(x) − ψ(y) and (ψ(x) − ψ(y))2 for certain values of
x and y, J. Comput. Appl. Math. (37) 1–3 (1991), 125–141.
 Book Review 113

[6] P. Flajolet, B. Salvy: Euler sums and contour integral representations, Exp. Math. (7) 1 (1998),
15–35.
[7] O. Furdui: Limits, Series, and Fractional Part Integrals. Problems in Mathematical Analysis,
Problem Books in Mathematics, Springer, 2013.
[8] H. F. Sandham: Problem 4305, Amer. Math. Monthly (55) 7 (1948), 431.
[9] F. A. Sherk, P. McMullen, A. C. Thompson, A. I. Weiss: Kaleidoscopes: Selected Writings of H.
S. M. Coxeter, Canadian Mathematical Society Series of Monographs and Advanced Texts, John
Wiley and Sons, Inc., New York, 1995.
[10] H. M. Srivastava, J. Choi: Zeta and q-Zeta Functions and Associated Series and Integrals, Elsevier,
Amsterdam, 2012.
[11] C. I. Vălean: A master theorem of series and an evaluation of a cubic harmonic series, J. Class.
Anal. (10) 2 (2017), 97–107.
[12] J. Wolstenholme: A Book of Mathematical Problems, on Subjects Included in the Cambridge
Course, MacMillian, London, 1867.

Seán M. Stewart After leaving Australia, for many years Seán taught mathematics to
engineers in Kazakhstan and the United Arab Emirates. He has always found it hard to resist
the challenge of a definite integral and is the author of How to integrate it: A practical guide to
finding elementary integrals published by Cambridge University Press.
9 Tanang Street, Bomaderry NSW 2541, Australia
E-mail address: [email protected]