Documentos de Académico
Documentos de Profesional
Documentos de Cultura
AND
DEVELOPMENT OF CALCULUS
M.D.SRINIVAS
CENTRE FOR POLICY STUDIES
mdsrinivas50@gmail.com
[ - ]
Bhskara, it seems, had not fully mastered this kind of "calculation with
infinitesimals" as is clear from some of the examples he considers in
Bjagaita, while solving quadratic equations by eliminating the middle
term (ekavara-madhyamharaa).
[ ]
[ .- ]
[ . ]
student
of
Nlakaha:
Karamta,
...
"Why then has an approximate value been mentioned here instead of the
actual value? This is the explanation. Because the actual value cannot be
expressed. Why? Given a certain unit of measurement in which the
diameter has no fractional part, the same measure when applied to
measure the circumference will certainly have a fractional part. ...Thus
when both are measured by the same unit they cannot both without
fractional parts. Even if you go a long way (by choosing smaller and
smaller units of measure) a small fractional part will remain. The import
[of sanna] is that there will never be a situation where both are integral."
Now substituting for (b-c)/b on the right from the equation and iterating
we get after m
Sn(k)=1k+2k++nk nk+1/(k+1)forlargen
They also give an estimate for the repeated summation (vra-sakalita)
Vn(1)=1+2+3+...+n=n(n+1)/2
Vn(k)=V1(k1)+V2(k1)++Vn(k1)nk+1/[(k+1)!]forlargen
C=4d[11/3+....+(1)(p1)/21/p+...]
In fact, the famous verses of Mdhava, which give the relation between
the circumference and diameter, also include the end-correction term
C=4d[11/3+....+...+(1)(p1)/21/p
+(1)(p+1)/2{(p+1)/2}/{(p+1)2+1}]
C=4d[11/3+....+...+(1)(p1)/21/p
+(1)(p+1)/2[{(p+1)/2}2+1]/[{(p+1)2+5}{(p+1)/2}]
To Mdhava is attributed a value of accurate to eleven decimal places
which is obtained by just computing fifty terms with the above correction.
Using the above correction term for p = 1, we get what may be called the
Mdhava continued fraction for :
2/(4)=2+12/2+22/2+32/2+...
A HISTORY OF APPROXIMATIONS TO
Approximation to
Accuracy
(Decimal
places)
1
Method Adopted
256/81 = 3.1604
Geometrical
25/8 = 3.125
Geometrical
3.0883
Geometrical
(10) = 3.1623
Geometrical
3 17/120 = 3. 141666
3.14159
Polygon doubling
(6.24 = 96 sides)
Polygon doubling
(6.26 = 384 sides)
Polygon doubling
(6.29 = 3072 sides)
Tsu Chhung-Chih
(480?)
355/113 = 3.1415929
3.1415927
6
7
ryabhaa (499)
62832/20000 = 3.1416
Polygon doubling
(6.29 = 12288
sides)
Polygon doubling
(4.28 = 1024 sides)
A HISTORY OF APPROXIMATIONS TO
Approximation to
Mdhava (1375)
Al Kashi (1430)
Francois Viete (1579)
Romanus (1593)
Ludolph Van Ceulen
(1615)
Wildebrod Snell
(1621)
Grienberger (1630)
Isaac Newton (1665)
Abraham Sharp (1699)
John Machin (1706)
Ramanujan (1910,
1914) Gosper (1985)
Kondo, Yee (2010)
2827433388233/9.1011
= 3.141592653592
3.1415926535897932
3.1415926536
3.1415926535.....
3.1415926535.....
3.1415926535.....
3.1415926535.....
3.1415926535.....
3.1415926535.....
3.1415926535.....
Ramanujan also notes that the last series "is extremely rapidly convergent".
Indeed in late 1980s, it blazed a new trail in the saga of computation of
Let M j+1 be the mid-point of the arc-bit CjCj+1 and similarly M j the midpoint of the previous (j-th) arc-bit. We shall denote the pia-jy of the arc
EMj+1 as Bj+1/2 and clearly Bj+1/2 = Mj+1Qj+1. The corresponding
Kj+1/2 = Mj+1Uj+1 and Sj+1/2 = EQj+1. Similarly, Bj-1/2 = MjQj, Kj-1/2 = MjUj and
Sj-1/2 = EQj. The full-chord of the arc-bit s/n may be denoted . Then a
simple argument based on trairika (similar triangles) leads to the
relations for Rsine and Rcosine differences
j=Bj+1Bj=(/R)Kj+1/2
Kj1/2Kj+1/2=(Sj+1/2Sj1/2)=(/R)Bj
Hence we have the ryabhaarelation for the second order differences in
the form derived by Nlakaha
j+1j = (Bj+1Bj)(BjBj1)=(/R)2Bj=(21)Bj/B
Yuktibhasa and Yuktidipika give the derivation of this and the Rversine
series by dividing the arc s into a large number n of equal parts and using
the relations between second-order Rsine differences (khaajyntara) and
the Rsines of arcs ns/j (piajys)
The verses giving the Rsine and Rversine series also note that the method
of obtaining accurate approximations to Rsine and Rversine values as
encoded in the mnemonics (also due to Madhava) Vidvn etc and Stena
etc, indeed follow from these series.
=m+Sin1[(r0/R)(1/R)Rsin(m)]
Nlakaha gives the correct formula for the correction to the mean
velocity in his treatise Tantrasagraha.
Nlakaha gives the derivative of the second term above in the form
[{(r0/R)Rcos(m)}/{R2(r0/R)2Rsin2(m)}1/2][(d/dt)(m)]