Srinivasaramanujan
Srinivasaramanujan
Srinivasaramanujan
bakonam), Indian mathematician whose contributions to the theory of numbers include pioneer-
ing discoveries of the properties of the partition function.
The story of the number 1729 goes back to 1918 when Indian mathematician Srinivasa Ramanu-
jan lay sick in a clinic near London and his friend and collaborator G.H. Hardy paid him a visit.
Hardy said that he had arrived in taxi number 1729 and described the number “as rather a dull
one.” Ramanujan replied to that saying, “No, Hardy, it’s a very interesting number! It’s the
smallest number expressible as the sum of two cubes in two different ways.”
Ramanujan, in his ailing state saw that 1729 can be represented as:
13 + 123 = 1729 = 93 + 103 .
Ta(n): the least positive integer that can be written as sum of two cubes in n different ways.
In 1938, G. H. Hardy and E. M. Wright proved that such numbers exist for all positive integers n,
and their proof is easily converted into a program to generate such numbers. However, the proof
makes no claims at all about whether the thus-generated numbers are the smallest possible and
thus it cannot be used to find the actual value of Ta(n).
The taxicab numbers subsequent to 1729 were found with the help of computers. Ta(3) was
found in 1957, Ta(4) in 1989, Ta(5) in 1994 and Ta(6) was announced on March 9, 2008. To
date, only six taxi-cab numbers have been discovered.
1
Ta(1) = 2 = 13 + 13
Ta(2) = 1729 = 13 + 123 = 93 + 103
Ta(3) = 87539319 = 1673 + 4363 = 2283 + 4233 = 2553 + 4143
Ta(4) = 6963472309248 = 24213 + 190833 = 54363 + 189483
= 102003 + 180723 = 133223 + 166303
Ta(5) = 48988659276962496 = 387873 + 3657573 = 1078393 + 3627533
= 2052923 + 3429523 = 2214243 + 3365883
= 2315183 + 3319543
Ta(6) = 24153319581254312065344 = 5821623 + 289062063 = 30641733 + 288948033
= 85192813 + 286574873 = 162180683 + 270932083
= 174924963 + 265904523 = 182899223 + 262243663
A more restrictive taxicab problem requires that the taxicab number be cube-free, which means
that it is not divisible by any cube other than 13 . When a cube-free taxicab number T is written
as T = x3 + y 3 , the numbers x and y must be relatively prime. Among the taxicab numbers
Ta(n) listed above, only Ta(1) and Ta(2) are cube-free taxicab numbers. The smallest cube-free
taxicab number with three representations was discovered in 1981. It is
15170835645 = 5173 + 24683 = 7093 + 24563 = 17333 + 21523 .
The smallest cube-free taxicab number with four representations was discovered in 2003. It is
https://en.wikipedia.org/wiki/Taxicab_number
2
Bronze Bust Ramanujan Statue, Rs 2.25 Lakhs / Piece
https://www.indiamart.com/shirgaonkarstudio/human-half-bust-statue.html
3
Volume One of
THE WORLD OF
MATHEMATICS
A small library of the literature
of mathematics from Ath-mose
the Scribe to Albert Einstein,
presented with commentaries and
notes by JAMES R. NEWMAN
flo-
-~
5
COMMENTARY ON
SRINIVASA RAMANUJAN
HA VE here set down from the scanty materials available, a brief ac-
I count of the poor Indian boy who became, as one eminent authority
has written, "quite the most extraordinary mathematician of our time."
Srinivasa Ramanujan died in India of tuberculosis on April 26, 1920, at
the age of thirty-three. He was a mathematician whom only first-class
mathematicians can follow and it is not surprising, therefore, that he at-
tracted little attention outside his profession. But his work has left a
memorable imprint on mathematical thought.
Two points provide the background for this sketch. The first is that,
despite a very limited formal education, Ramanujan was already a brilliant
mathematician when he came to England to study in 1914. On the founda-
tion of a borrowed volume, Carr's Synopsis of Pure Mathematics, he had
built "an outstanding edifice of analytical knowledge and discovery." 1 It
was the only book on higher mathematics to which Ramanujan had access,
and the nature of his achievement becomes clear when one examines
Carr's text. While a work of "some real scholarship and enthusiasm and
with a style and individuality of its own," it was in fact no more than a
synopsis of some 6,000 theorems of algebra, trigonometry, calculus and
analytical geometry with proofs "which are often little more than cross
references." :! In general, the mathematical knowledge contained in Carr's
book went no further than the 1860s. Yet in areas that interested him,
Ramanujan was abreast, and often ahead, of contemporary mathematical
knowledge when he arrived in England. Thus in a mighty sweep he had
succeeded in re-creating in his field through his own unaided powers, a rich
half-century of European mathematics. One may doubt that so prodigious
a feat had ever before been accomplished in the history of thought.
The second noteworthy point is that Ramanujan was a particular kind
of mathematician. He was not as versatile as Gauss or Poincare. He was
not a geometer; he cared nothing for mathematical physics, let alone the
possible "usefulness" of his mathematical work to other disciplines. In-
stead, Ramanujan's intuition was much at ease in the bewildering inter-
1 The book was A Synopsif of Elememary Re~lllts ill Pllre and Applied Mathe·
matics, by George Shoobridge Carr, a Cambridge mathematician. It was published
in two volumes, 1880 and 1886. Carr waS a private coach in London and came to
Cambridge as an undergraduate when he was nearly forty. The book is "substantially
a summary of Carr's coaching notes . . . . He is now completely forgotten, even in
his own college, except in so far as Ramanujan has kept his name alive, but he must
have been in some ways rather a remarkable man." G. H. Hardy. Ramallujall-Twe/l'e
Lectures Suggested by His Life and Work~ Cambridge. 1940, p. 3.
2 G. H. Hardy, op. cit., p. 3.
366
S,lnlvaSil Ramanu/an 367
stices of the number system. Numbers, as will appear, were his friends. In
the simplest array of digits he detected wonderful properties: congruences,
symmetries and relationships which had escaped the notice of even the
outstandingly gifted theoreticians. The modem theory of numbers is at
once one of the richest, most elusive and most difficult branches of mathe-
matics. Some of its principa1 theorems. while self-evident and childishly
simple in statement, defy repeated and strenuous efforts to prove them. A
good example is G01dbach's Theorem, which states that every even num-
ber is the sum of two prime numbers. Any fool, as one noted mathema-
tician remarked, might have thought of it; it is altogether obvious, and no
even number has been found which does not obey it. Yet no proof of its
validity for every even number has yet been adduced. It was in dealing
with such problems that Ramanujan showed his remarkable powers.
The 1ate G. H. Hardy, a leading mathematician of his time (see pp. 2024-
2026), was professionally and personally closest to Ramanujan during his
fruitful five years in England. I have taken from Hardy's wen-known
obituary of Ramanujan,:-t and from his notable co~rse of Ramanujan lec-
tures at Harvard 4 the bulk of the material in the selection following; the
rest comes from a brief biographical sketch by P. V. Seshu Aiyar and
R. Ramachandra Rao to be found in Ramanujan's Collected Works.;) My
contribution has been merely to copy, paraphrase and select. Some of the
materjal is understandable only to the professional mathematician. There
is enough, I think, of general interest to justify bringing before the com-
mon reader even this inadequate notice of a true genius.
aproceedingsoflhe London MalhematicalSociely (2), XIX (1921), pp. XL-LVIII.
Reprinted in Collected Papers 01 Srini~'asa Ramam.jatr, edited by G. H. Hardy, P. V.
Seshu Aiyar and B. M. Wilson, Cambridge, 1927, pp. XXI-XXXVI.
.. G. H. Hardy, Ramanujan-Twelve Lectures Suggested by His Life and Work:
Cambridge, 1940.
~ See footnote 3, above.
1 have often admired the mystical way 0/ Pythagoras and the secret magic
0/ numbers. -Sill THOMAS BROWNE
13 Srinivasa Ramal1ujan
By JAMES R. NEWMAN
1911, when Ramanujan was 23. His first long article was on "Some Prop-
erties of Bernoulli's Numbers" and was published in the same year. In
1912 he contributed two more notes to the same journal and also several
questions for solution.
By this time Ramachandra Rao had induced a Mr. Griffith of the Ma-
dras Engineering College to take an interest in Ramanujan, and Griffith
spoke to Sir Francis Spring, the chairman of the Madras Port Trust, where
Ramanujan was employed. From that time on it became easy to secure
recognition of his work. Upon the suggestion of Seshu Aiyar and others,
Ramanujan began a correspondence with G. H. Hardy, then Fellow of
Trinity College, Cambridge. His first letter to Hardy, dated January 16,
1913, which his friends helped him put in English, fol1ows:
UDEAR 8m,
"I beg to introduce myself to you as a clerk in the Accounts Depart-
ment of the Port Trust Office at Madras on a salary of only £20 per
annum. I am now about 23 years of age. [He was actually 25-Ed.] I have
had no University education but I have undergone the ordinary school
course. After leaving school I have been employing the spare time at my
disposal to work at Mathematics. I have not trodden through the conven-
tional regular course which is followed in a University course, but I am
striking out a new path for myself. I have made a special investigation of
divergent series in general and the results I get are termed by the local
mathematicians as 'startling'. . . .
"I would request you to go through the enclosed papers. Being poor, if
you are convinced that there is anything of value I would like to have my
theorems published. I have not given the actual investigations nor the
expressions that I get but I have indicated the lines on which I proceed.
Being inexperienced I would very highly va] ue any advice you give me.
Requesting to be excused for the trouble I give you.
"I remain, Dear Sir, Yours truly,
"S. RAMANUJAN."
To the letter were attached about 120 theorems, of which the 15 here
presented were part of a group selected by Hardy as "fairly representa·
tive. n Hardy commented on these:
uI should like you to begin by trying to reconstruct the immediate re-
31 6!
(1.1) 1- x2 + x4 - ...
(1 !2!)3 (2!4!)3
Sriniwua RGlfltUluJfIII 371
( 1.2) 1_ 5 (!.)32
+ 9 ( 1.
2.4
3)3 _13 ( 1.3 .5)3+ ... = :.
2.4.6 T
(1.4) 1 - 5 -
1):;
(2 2 .4
+
(1.3):; 13 (1.3.5):'
9 - + ...
- =
2 .
2 .4 . 6 {r( % ) l'
00
1+ (_X)2b+1
1+(
b+2
X)2
( 1.5) - - - - - ... dx
o 1+(~r I+C:S
r( a + ¥,,) r( b + 1) r( b - a + 1),)
=%~~--------------------~
r( a) r( b %) r( b - a + 1) +
(1.6)/00 _ _ _-_d_x_ _ _ __
o (1 + x2) (l + r2x2)( 1 + r4x 2 ) . ••
a-~ 1 + 4a
00 xe-
aZ2
dx
)
= IJ-~
(
1 + 41J
/00 xe-/U'2 )
dx .
( / /I e21TJ _ 1 0 e2frZ - 1
a 2 3 e-a2 1 4
( 1.8) e-.r dx = 1f.z~lh - - - - - - - -
2
.
/ o 2a+a+2a+a+2a+ ...
-1-+-V'-!-S~-'(-~-:-2-1-)-%-1-1- V + ]e2
5
r/
2 "'.
KJ:r dt + fJ(x),
..I. y(log t)
where K = 0·764 ... and fJ(x) is very small compared with the previous
integral.
theorems and was keeping a great deal up his sleeve. The second is a
formula of Bauer well known in the theory of Legendre series, but the
others are much harder than they look. . , ,
·'The formulae (1.10) -( 1.13) are on a different level and obviously
both difficult and deep. An expert in elliptic functions can see at once that
0.13) is derived somehow from the theory of 'complex multiplication;
but (1.10)-(1.12) defeated me completely; I had never seen anything in
the least like them before. A single look at them is enough to show that
they could only be written down by a mathematician of the highest class.
They must be true because, if they were not true, no one would have had
the imagination to invent them. Finally . . . the writer must be com-
pletely honest, because great mathematicians are commoner than thieves
or humbugs of such incredible skill. . . .
"While Ramanujan had numerous brilliant successes, his work on prime
numbers and on all the allied problems of the theory was definitely wrong.
This may be said to have been his one great failure. And yet I am not sure
that, in some ways, his failure was not more wonderful than any of his
triumphs. . . . n
Ramanujan's notation of one mathematical term in this area, wrote
Hardy, "was first obtained by Landau in 1908. Ramanujan had none of
Landau's weapons at his command; he had never seen a French or Ger-
man book; his knowledge even of English was insufficient to qualify for a
degree. It is sufficiently marvellous that he should have even dreamt of
problems such as these, problems which it had taken the finest mathe-
maticians in Europe a hundred years to solve, and of which the solution
is incomplete to the present day."
At last, in May of 1913, as the result of the help of many friends,
Ramanujan was relieved of his clerical post in the Madras Port Trust and
given a special scholarship. Hardy had made efforts from the first to bring
Ramanujan to Cambridge. The way seemed to be open, but Ramanujan
refused at first to go because of caste prejudice and lack of his mother's
consent.
"This consent," wrote Hardy, "was at last got very easily in an unex-
pected manner. For one morning his mother announced that she had had
a dream on the previous night, in which she saw her son seated in a big
hall amidst a group of Europeans, and that the goddess Namagiri had
commanded her not to stand in the way of her son fulfilling his life's
purpose."
When Ramanujan finally came, he had a scholarship from Madras of
£250, of which £50 was allotted to the support of his family in India, and
an allowance of £60 from Trinity,
'There was one great puzzle," Hardy observes of Ramanujan. "What was
to be done in the way of teaching him modern mathematics? The limita-
The Man who knew infinity a biographical movie on Ramanujan was released on August 2016:
https://www.youtube.com/watch?v=hlXHwMgS06c
14
The movie is based on the book:
The Man Who Knew Infinity: A Life of the Genius Ramanujan,
a biography of Srinivasa Ramanujan, written in 1991 by Robert Kanigel and published by Wash-
ington Square Press.
15
Ramanujan’s Notebooks
The history of the notebooks, in brief, is the following: Ramanujan had noted down the results
of his researches, without proofs, (as in A Synopsis of Elementary Results, a book on pure
Mathematics, by G.S. Carr), in three notebooks, between the years 1903 - 1914, before he left
for England. These were the notebooks which he showed to his benefactors to convince them
about his abilities as a mathematician. The results in these notebooks were organized by him.
The first notebook has 16 chapters in 134 pages. The second is a revised, enlarged version of the
first, containing 21 chapters in 252 pages. The third notebook contains 33 pages of unorganized
material. Ramanujan took these notebooks with him to Cambridge. But, in one of his letters to a
friend, he wrote that he had no time to look into them and most probably he did not put them to
use during his five year stay abroad.
Carr’s Book
Hardy tried to persuade the University of Madras to undertake such a task and in turn, Prof. G.N.
Watson (31 January 1886 – 2 February 1965) was requested by the Madras University, in 1931,
to edit the notebooks in a suitable form for publication. This was a formidable task since the
notebooks contained about 4000 theorems.
The formidable task of truly editing the notebooks - viz. to either prove each of the results or
provide references to the literature where the proofs may be found - which was started by G. N.
16
Watson and B. M. Wilson but never completed, was taken up in right earnest by Prof. Bruce C.
Berndt (b. March 13, 1939). After publishing series of papers with various collaborators, since
1981, he finally published five parts from Springer Publications:
Ramanujan’s Notebooks, Part I (1985), Part II (1989), Part III (1991), Part IV (1994), Part V
(1998).
17
Ramanujan Notebook
18
Ramanujan’s lost notebook
Prof. Watson died in Feb. 1965. In the spring of 1976, when Prof. George Andrews of the
Pennsylvania State University visited Trinity College Library at Cambridge University. It was
suggested to him that there were materials deposited there from the estate of the late Prof. G. N.
Watson that might be of interest to him. In one box of materials from Watson’s estate, Andrews
found several items written by Ramanujan, with the most interesting item being a manuscript
written on 138 sides in Ramanujan’s distinctive handwriting. The sheets contained over six
hundred formulas without proofs. It is assumed that this manuscript, or notebook, was written
during the last year of Ramanujan’s life after his return to India from England. These are now
referred to as the contents of the Lost notebook of Ramanujan.
George Andrews (b. 4 December 1938) and Bruce C. Berndt have published books in five parts
in which they give proofs for Ramanujan’s formulas included in the notebook.
Ramanujan’s Lost Notebook (Springer) Part I (2005), Part II (2009), Part III (2012), Part IV
(2013), Part V (2018).
19
Ramanujan and the squaring of circle.
Squaring the circle is a problem proposed by ancient geometers. It is the challenge of construct-
ing a square with the same area as a given circle by using only a finite number of steps with
compass and straightedge.
This was proven to be impossible, as a consequence of the Lindemann–Weierstrass theorem in
1882.
The Rhind Papyrus: dated 1550 BC, written by a scribe named Ahmes, gave the approximate
2
16
value of for π. (i.e. 3.16049)
9
355
In 1913 Ramanujan gave construction that shows π = .
113
Ramanujan 1914
20
5
SQUARING THE CIRCLE
(Journal of the Indian Matltematical Societ!!, v, 1913, 132)
Let PQR be a circle with centre 0, of which a diameter is P R. Biseet
PO at H and let 1.' be the point of trisection of OR nearer R. Draw TQ
perpendicular to P R and place the chord RS = TQ.
Join PS,and draw OM and TN parallel to RS. Place a chord PK = PM,
and draw the tangent P L = MN. Join RL. RK and KL. Cut off RO = RH.
Draw OD parallel to KL, meeting RL at D.
'rhen the square on RD will be equal to the circle PQR approximately.
For RS' = :fa d2,
where d is the diameter of the circle.
'l'herefore
But PL and PK are equal to J.1IN and PM respectively.
Therefore Pl(' = N"d', ,UlU PL" = -j,hd2.
Hence R](2 = PR' - P](!J = tHd',
and RL' = P B" + P L' = llid·.
RK _ RG _ :1 /113
But
RL -lW- 2\1 355'
and lW=td.
d /355
Therefore RD = 2 \I I 1:3 = r ..j7T", very nearly.
N Qte.-If the urea of the circle be 140,000 square miles, then RD is
greater than the true length by about an inch.
240
Srinivasa Ramanujan in 1914 gave a ruler-and-compass construction which was equivalent to
taking the approximate value for π to be
!1 s
2 192 4
4 2143
9 + = = 3.1415926525826461252 . . . giving eight decimal places of π.
22 22
22
With the development of computer technology in the 1950s, π was computed to thousands and
then millions of digits. These computations were facilitated by the discovery that high- precision
multiplication could be performed rapidly using fast Fourier transform (FFT)s. In spite of these
advances, until the 1970s all computer evaluations of π still employed classical formulas, usually
one of the Machin-type formulas.
π 1 1
= 4 tan−1 − tan−1 .
4 5 239
(John Machin (1686-1751) in 1706).
π 1 1 1 1
= 4× − + − + ···
4 5 3 · 53 5 · 55 7 · 57
1 1 1 1
− − + − + · · ·
239 3 · 2393 5 · 2395 7 · 2397
In 1949 George Reitwiesner managed to calculate 2037 digits of π using the first electronic
computer ENIAC (Electronic Numerical Integrator and Computer) in 70 hours. He used the
following Machin like expression given by Ferguson in 1945.
π 1 1 1
= 3 tan−1 + tan−1 + tan−1 .
4 4 20 1985
The 2002 record for digits of π, 1,241,100,000,000, was obtained by Yasumasa Kanada of Tokyo
University. The calculation was performed on a 64-node Hitachi supercomputer with 1 terabyte
of main memory, performing 2 trillion operations per second. It took 600 hours to complete this
task. The following two equations were both used:
π 1 1 1 1
= 12 arctan + 32 arctan − 5 arctan + 12 arctan
4 49 57 239 110443
Kikuo Takano (1982)
π 1 1 1 1
= 44 arctan + 7 arctan − 12 arctan + 24 arctan
4 57 239 682 12943
F. C. M. Stormer (1896)
23
Yamasa Kamada
Until 1970 most π calculators were using variations of Machin’s formula. Some new infinite
series formulas were discovered by Ramanujan around 1910, but these were not well known
until quite recently when his writings were widely published. In 1980-90 some new rapidly
converging infinite series for π were discovered. Most significant one was a series given by
Srinivasa Ramanujan (1887-1920) in 1914.
√ ∞
1 2 2 X (4n)!(1103 + 26390n)
= .
π 9801 n=0 (n!)4 3964n
9801
• The first term of this series gives π ≈ √
2206 2
≈ 3.1415927; this is π correct upto six
decimal places!
• Each term of this series produces an additional 8 correct digits in the result.
• This series was used by William Gosper to compute 17 million digits of π in 1985.
At about the same time, Chudnovsky brothers (David and Gregory) found the following variation
of Ramanujan’s formula:
∞
1 X (−1)n (6n)!(13591409 + 545140134n)
= 12 .
π n=0 (3n)!(n!)3 6403203n+3/2
Each term of this series produces an additional 14 correct digits. The Chudnovskys implemented
this formula using a clever scheme that enabled them to utilize the results of an initial level of
precision to extend the calculation to even higher precision. They used this method in several
large calculations of π, culminating with a computation to over four billion decimal digits in
1994.
24
Chudnovsky brothers
In 1987 Ramanujam series was improved to develop a faster converging that gives 14 digits
of π per term, and has been used for several record-setting π calculations, including the first
to surpass 1 billion (1,011,196,691) digits in 1989 by the Chudnovsky brothers, 2.7 trillion
(2,699,999,990,000) digits by Fabrice Bellard in 2009, and 10 trillion (10,000,000,000,050) dig-
its in 2011 by Alexander Yee and Shigeru Kondo.
Further details can be seen
https://en.wikipedia.org/wiki/Chudnovsky_algorithm
https://en.wikipedia.org/wiki/Borwein%27s_algorithm
25
Name Year Correct Digits
Archimedes 250? bce 3
Ptolemy 150 3
Liu Hui 263 5
Tsu Ch’ung Chi 480? 7
Al-Kashi 1429 14
Romanus 1593 15
Van Ceulen 1615 35
Sharp 1699 71
Machin 1706 100
Strassnitzky and Dase 1844 200
Rutherford 1853 440
Shanks 1874 527
Reitwiesner et al. (ENIAC) 1949 2,037
Genuys 1958 10,000
Shanks and Wrench 1961 100,265
Guilloud and Bouyer 1973 1,001,250
Miyoshi and Kanada 1981 2,000,036
Kanada, Yoshino and Tamura 1982 16,777,206
Gosper 1985 17,526,200
Bailey Jan. 1986 29,360,111
Kanada and Tamura Sep. 1986 33,554,414
Kanada and Tamura Oct. 1986 67,108,839
Kanada et. al Jan. 1987 134,217,700
Kanada and Tamura Jan. 1988 201,326,551
Chudnovskys May 1989 480,000,000
Kanada and Tamura Jul. 1989 536,870,898
Kanada and Tamura Nov. 1989 1,073,741,799
Chudnovskys Aug. 1991 2,260,000,000
Chudnovskys May 1994 4,044,000,000
Kanada and Takahashi Oct. 1995 6,442,450,938
Kanada and Takahashi Jul. 1997 51,539,600,000
Kanada and Takahashi Sep. 1999 206,158,430,000
Kanada, Ushiro, Kuroda Dec. 2002 1,241,100,000,000
4
Pi has been calculated out to 31.4 trillion decimals, Google announces on Pi Day, 2019.
Google revealed developer advocate Emma Haruka Iwao, with the help of the tech giant’s cloud
platform, calculated Pi to 31.4 trillion decimal places, beating the previous record by nearly 9
trillion digits.
The calculation required 170 terabytes of data, about the same amount of data as the entire
Library of Congress print collection, said Google.
https://techxplore.com/news/2019-03-pi-trillion-decimals-google-day.html
https://www.youtube.com/watch?time_continue=132&v=JvEvTcXF-4Q&feature=
emb_logo
27
2004, Springer Science+Business Media New York
28
https://www.washingtonpost.com/news/wonk/wp/2015/03/14/10-stunning-images-show-the-beauty-hidden-in-pi/
29
This is a walk made out of the first 100 billion digits of pi in base 4 with the following rules for
the steps: 0 right, 1 up, 2 left, 3 down.
30
Kerala Mathematics for computation of pi
C D
R
G
S
K H
A B
The proof in Jyestadeva’s book is by direct rectification of an arc of a circle. In the diagram
below arc AD is a quarter circle of radius 1, center at C and ABCD is a square. the side BD
is divided in n equal parts each of length δ so that nδ = 1. Let GH = δ. RS and GK are
perpendicular to CH. Then triangles CRS and CGK are similar.
RS GK GK
= , that is, RS = .
RC CG CG
Similarity of triangles CHD and GHK imply that
GH GK GH
= , that is, GK = .
CH CD CH
Therefore
GH GH δ δ
RS = ' 2
= 2
= .
CG.CH CH 1 + DH 1 + r2 δ 2
Here GH is taken as the r-th section on BD.
δ
Since arc(RS) ' RS = , so the quarter of the circle is
1 + r2 δ 2
n
π X δ
= lim .
4 n→∞ r=1 1 + r2 δ 2
The clear idea of limit did not exist that time, so the relation is to be understood in the intuitive
sense only.
Now after this is more of intuition than mathematics.
31
For the geometric series
1
= 1 + x + x2 · · · ,
1+x
Jyesthadeva used the following iterative procedure:
1 1 1
=1−x =1−x 1−x .
1+x 1+x 1+x
np+1
Sn(p) := 1p + 2p + · · · + np ∼ for large n. (1)
p+1
Sn(1) = n + (n − 1) + (n − 2) + · · · = n2 − (1 + · · · + n − 1 + n) − n = n2 − n − Sn(1)
1 n2
(1) 2 2
=⇒ 2Sn = n − n = n 1 − ∼ n2 or Sn(1) ∼
n 2
n(n + 1) n2 1 n2
Sn(1) = 1 + · · · + n = = 1+ ∼ for large n.
2 2 n 2
32
Then for large n:
33
If not exact we can minimize the error term Er(n) = f (n + 1) + f (n − 1) − n1 . It is clear that
1
f (n) should be of O( n1 ). Otherwise Er(n) will blow up. We assume that f (n) = . Thus
an + b
1 1 1 1
Er(n) = f (n + 1) + f (n − 1) − = + −
n a(n + 1) + b a(n − 1) + b n
(2a − a )n + 2b(1 − a)n + a2 + b2
2 2
=
n((an + b)2 − a2 )
Thus for least we should have no n2 and n terms, that is, a = 2, b = 0. In this case
1 1
f (n) = , Er(n) = .
2n n3 − n
This estimate of the inaccuracy, Er(n) being positive, shows that the correction has been over
done and hence there has to be a reduction in the correction. This means that f (n) is to be
1
decreased. If we take f (n) = 2n+1 , then
1 1 1 −2n + 3
Er(n) = + − = 3 ∼ O(n−2 ).
2n + 3 2n + 1 n 4n + 4n2 − 3n
Jyesthadeva proposed that as the correction in the denominator of f (n) is less than 1, we take the
c
denominator of f (n) as 2n + 2n , where c is to be determined. Thus new
1
f (n) = .
c
2n +
2n
Again with this f (n):
1 2(n + 1) 2(n − 1) 1
Er(n) = f (n + 1) + f (n − 1) − = + −
n 4(n + 1) + c 4(n − 1) + c n
2 2
2 2
(16 − 4c)n − (c + 4)
=
n(4(n + 1)2 + c)(4(n − 1)2 + c)
−4
Thus if c = 4 we have Er(n) = ∼ O(n−5 ). In this case
n5 + 4n
1 n/2
f (n) = = .
4 n2 + 1
2n +
2n
34
Let us call this the second correction, denote by:
n/2
f2 (n) = . (3)
n2 + 1
Similar next correction f3 is
1
.
4
2n +
16
2n +
2n
In fact,
n2 /4 + 1
f3 (x) = (4)
(n2 + 5)n/2
The error is
2304
Er(n) = ∼ o(n−7 ).
64n7 + 448n5 + 1792n3 − 2304n
The next correction is
1
.
22
2n +
42
2n +
62
2n +
2n
The third correction is very effective in obtaining a good numerical value of π without much
calculation. For example
1 1
1 − + ··· − + f3 (20)
3 19
gives value of π correct to eight decimal places. Nilkantha himself gives 104348/33215 which
is correct up to nine decimal places.
Next this transforms the series for π/4 to a faster convergence series.
π 1 1 1 1
= 1 − + − + − ···
4 3 5 7 9
1 1 1
= (1 − f (2)) + (f (2) + f (4) − ) − (f (4) + f (6) − ) + (f (6) + f (8) − ) − · · ·
3 5 7
= (1 − f (2)) + Er(3) − Er(5) + Er(7) − Er(9) + · · ·
1 1
If f (n) = , then Er(n) = 3 , and the series is
2n n −n
π 1 1 1 1 1
=1− + 3 − 3 + 3 − 3 + ···
4 4 3 −3 5 −5 7 −7 9 −9
35
n/2 −4
For the other correction f (n) = , we have Er(n) = . Thus
n2 + 1 n5 + 4n
π 1 4 4 4
=1− − 5 + 5 − 5 + ···
4 5 3 + 4.3 5 + 4.5 7 + 4.7
4 4 4 4
= 5 − 5 + 5 − 5 + ···
1 + 4.1 3 + 4.3 5 + 4.5 7 + 4.7
36