Upload
others
View
7
Download
0
Embed Size (px)
Citation preview
SUMMATION OF DIVERGENT INFINITE SERIES BY ARITHMETIC, GEOMETRIC, AND HARMONIC
MEANS
by
FREDERICK HARRIS YOUNG
A THESIS
submitted to
OREGON STATE COLLEGE
In partial t'u1fi11nnt of the requirements for the
degree of
MASTER OF ARTS
Juno j913
APPROVED:
Professor of Department of Mathematics
In Charge of Major
Head of Department of )4athematics
Chairman of School Graduate Comxuitte
Dean of Graduate Sohool
ACKNOWLEDGMENT
The writer wishes to express his deep gratitude
to Dr. C. L. Clark for the latter's skillful direction,
timely help, and abundant patience.
TABLE OF CONTENTS
Chapter Titlo
I. INTROI5ITCTION. . 1 , . . s., . , . , . . . .1
II TYPE S OF REQULA.R SUNMATI ON
À. Cesàr'o Summation.... , * . ...... . ... .3
B. aölder Suat ton. . . . . . . . . . . . . . . . . . . . b
C. Nörilund Mean. . s . s s a . * . a . . . a .9
III. HARMONIC-GEOMETRIC-ARITHMETIC EAN INEQUALITY. . . . . . . . . . . . . .. . . I. .10
IV. INTEGRAL I1EQJALIT).. . . . . . . a a . . . . . . . . . . .1)
V. SUNATIONBYIIâRMOXUCNEÀL...............ILI
VI. WEIGHTEDiiARMONICMEAN....................21
VII. ZUMMATIONBYGEOMETRICKEAN...............22
VIII. WEIGHPEDGEOMETRICMEAL..................2k
Ix. DIVERGENT INFINITE PRODUCPS...............25
X. FURTHER EXAMPLES OF HARMONIC MEAN
XIs COMPARISON OF PROPERTIES C? CESARO ui HARMONIC MEANS. . . . . . . . . . . . . . . . . . . . . . . . . .29
SUMMATION OF DIVERGENT INFINITE SERIES BY ARITHMETIC, GEOMETRIC, AND HARMONIC MEANS
I. INTROL1CTION
Prior to the time of Cauchy and Abel, infinite
series vere used with no regard for the question of con-
vergence. If a function could be expressed as an infinite
sortes, well and good; the eones represented the function.
This carefree attitude, which led to many discoveries
later validated by sounder means, was brought to a sudden
halt when Cauchy and Abel showed that the use of divergent
series could lead to incorrect results. Thus, with one
sweep, divergent series were relegated to the limbo of
things tried and found wanting. Not until ne8rly the end
of the nineteenth century did these treacherous traps for
the unwary begin to regain an aura of respectability.
Frobenius, in 1880, extended the idea of convergency by
employing the limit of the arithmetic mean of the partial
sums of a series instead of the limit of the sum of the
first n ternis.
The work of Frobenius opened the way for a host
of workers. A way had been found to attach a unique
vniue to certain nonconvergent series. Further, this
method, when applied to convergent series, yielded the
ordinary sum. Within a decade Hölder, Cesro, and Borel
had successfully generalized the method of Frobenius,
2
and the early twentieth eentuiy wiis a period of intürthe
and fruitful activity in suimnation.
Certain terms 8tloUld be defined at this point.
A "regular" method of summation
the ordinary value when applied
If, in addition, this method el;
when applied to a nonconvergent
known as a "conventional" sum.
infinite products.
ta one that will yield
to a convergent series. o yields a unique result
series, the result is
Similar terms apply to
The purpose of this paper is es follows:
i. To define two new methods of summation, those of
the harmonic mean and of the geometric mean of
the partial sums of a series, and to show these
methods to be regular.
2. To define a conventional product of an infinite
product as the geometric mean of the partial
products and to show this method is regular.
3. Po compere these methods of summation with those
of Ces.ro, Hölder, and Nörlund.
3
II. TYPE S OF REGULAR BUMMAPI ON
Cesro Summation
One of the moat useful methods of regular sum-
mation 1.3 that of Cesaro. cO (o) n
Consider the series u. Let . U. k=1
(i) n (o) (r) n (r-1) Let and, in general, let =
k=1
(r) (r) 1(r))
Let C rl(n - s] ; then, if for some r, CJ has n
tri + rfJT a defluite limit, the series
mable (C,r).
00
1 u1, is snid to be sum- k=1
S. Chapman and Knopp (5, p.70) have extended
Cesro's method to the case in which the order of summe-
bility, r, is not necessarily integral. Then
(r) (r) , where + n) denotes
(r+n n
(r + 11 jr + 2...(r4- n or fl(r + n + i) - - n. rCn + i7Çr 1f
( r ) (o? (p (o) /) + i.) (o) 1/r + n - 1\
and = s - 8n1 2 8n-2 " n Jo
For a proof of the regularity of Cesro sumnii.tion
(7, p.26k), let s = s and
cO A- uk=lime
k=1 n - -n:---
_!
Case 1. A (L Given any > O, there exists an
n > O such that (m + il' I + 21' < Hence, if
n
+ + .... + s i + {21 - vfl n - nl
= n n
Thus, !T fr-/E, orltmç=O. Case 2. A O. Then the sequence - tends
to zero, and ve my apply case I to it. That is,
n lin i (s, - A) = lim(cr - A) O;
n-'nk=1 n-'
therefore, lin A.
If we consider the points P, 2J ' which rcpreent the numbers s, 2' Sn'
w LA(.Y
happen that they do not converse o a single lixaitin
point. Howevei, the set of points , 2' n'
where is the centioid of , ... each wtth a masa
of unity, mny converge to !, which thsn represents the
Cesàro sum.
For an illustration of Cesèro's method, lot us
consider the series I - I + I - i + ... . Then
(o) (i) =1 32n
5
(o) (i) =0 22n+1n
(o)
52n =0
(o) 8 =1 2n + I
(i) (i) (i)
Then C2 n I C2 + ri and 11m C
2ñ+I n- 2n+1=.
Thus, the original series is summable (C,i) to the value
1/2.
For an illustration of Cesro summation of a series
not swnmable (C,i), let us consider the series
u = I(_l)kk. Then k=1
(o) (i)
si =_i si
=_I
(o) (i) = = O
(o) (i) S3 '=-2 83 =-2
(o) (i) = 2 = O
.........................
(o) (i)
$2n_1 82n-1
(o) (i)
82n = S - O
(2)
82n I -n(n + i)
-
(2)
82n -n(n + i)
3 (2)
21(2n - i - t) í-n' (n + i) f2n - I + 2- I7T )
(2) n +n ;limC -1 n-1=
2n
(2)
i-
(_)(n + i) 212r1
-
2n
(2) limC _-1.
2n -
Thus, the series U = (_i)kk is suinmable (C,2) to the k=1
value -rn. For another illustration, omehat different,
consider the series U = sin la where x is not an even k=i
multiple of 7. Then n
s = sin la cos x/2 - cosjn+í/21x, and n
k=1 - 2 sIr[x7 - n n
= s n cos x/2 I cos (k+1/2)x. n
k=1 n 2In c7 I1iE7 Ic=1
n Now, cos (k+1/2)x -sin x + sin (n+i)x. Therefore,
k=1 2sTn x/2
n cos x/2 sin x - sin (niJx flOE-. + iIn X7 ° that n - - 2sTh x/2
hm cos x/2 I cot x/2. n-*oo fl = 2iTh/2 =
However, if x is an even niultiple of r,
U = O ± O + ... and lin = O. Therefore, U is noo
sununable (c,i) to the value I cot x/2 if x is not an even
roi
L!J
Holder Sumnmt ion
H6lder's method is siniilar to that of Cesàro, and
the two methods may be shown to be equivalent (3, p.85).
n
Consider the series where s L. Uke k=1
(i) Let h + s +...+a, and, in general, let
(r) (r-1)(r-1) (r-1) h1 4 h2+... +h.
(r) If 11m h has r definite value for some integral
n -
value of r, the sertes 2 Uk is said to be suinmable
(H,r).
Since Hldor's method is equivrlent to that of
Cesro, it is necessarily regular.
9
Nórlund Meen
Consider the sequence fc1} oi' real or complex
consG1r1ts with C0 O, and let C c. For any k.o
00 ri sortes 1 u, where s Uk, and um n O, let
k=o k=o n- -
cl-j:1 CS + + + CrìSc S If li -
On "n 00
bae a definite value, the series - u U said to be k=o
summable (N;c).
The method of Nrlund's Mean can be shown to be a
recular method of summation (k, p.37).
It is interesting to note that if all c are i,
Nòrlund's iethod :Ls identical with those of Cesro and
Hölder for te first order,
III. HARMONIC-GEOMEPRIC-ARITBJIETIC MEAN INEQUALITY
A fundamentsl inequality which we shall find of
great use is the hrmonic-geometric-arithmetic mean in-
equality. The proof presented here is due to Cauchy
(i, p.)25). First, ve shall prove that the geometric mean of
several non-negative numbers A,B,C,D,... is always less
than their arithmetic mean.
Let n be the number of letters A,B,C,D,... . It
is sufficient to prove that in general
(i) BCD.Tt A+B+C+D+... n
or, what amounts to the same thing, that
(2) ABCD... -/A+B+C+D+ n
= (--- n -
In the f trat place, for n = 2,
+ B2 A - B A + B 2
- ___ - ( 2 ) = ( 2)
and one concludes, by taking successively n = , n = 8,
1 and finally n = 2m, that -
J ' J k
ABCflEFGH ________ ________ = - ¿ _//
¿
(A+B+c+D+E+F+a+H8, - 's___ _- - __-)
n
(3) ABCD... , (A + B + C + D + ...) - -- - 2m -_-- -J
In the second case, if n is not ' term of the
-
geometric progl'e8Siofl 2,l,8,16,..., let us designate by
a term of this progression larger than n, and we shell
let
K A + B + C + D + n
Then, returning to formula (3) and suppOsing in the first
member of this formula that the last 2m - n factors are
equal to K, we twve
ABCD...K2 LJA + B +C + L J
or, in other terme,
ABCD...K'
Then, multiplying both members by K' , we have
.ABCD... K (A+B C+D which was to be proved.
Nov, denote by A(a), O(a), and H(e) the arithmetic,
geometric, and harmonie means of a finite number (n) of
nonnegative numbers. Then, as n-* , we have
lin A(a) uni O(a), if the8e limita exist.
Also, for all a ) O, 'e know that A(1/a) = 1/H(a),
and i/a(e) = 0(1/a). Therefore, i/o() 0(1/a) A(1/a) =
1/il(e). Thus, 1/O(a) 1/H(s), or H(a) O(a), and
lin Hut) lin O(a), if these limits cxist. Therefore, - lin li(a) lin G(a) li i(a), if the limits n- n- n-*oo
exist.
i'.
If the 1imit of A(a), G(), and H(a) do not exist,
then the inequality holds true for the upper limite,
least upper bounds, etc. of the functions.
t)
IV. IIEGRAL I1EQ.UAIJITY
An interesting extension of the preceding in-
equality is the following: let us define, over the in-
terval O < ç X2,
A(f) i J f(x)dx, x2_xix1 X -(2
./ log f (x ) dx G(f)=eX21 X1
H(f) 2
- , where f(x) has a positive lower
dx I xl
bound. Then H(f) G(f) Á(f).
The general proof of this theoreni has been given by
Hardy, Littlewood, &ad Póly (2, p.135). However, the
inequality H(f) A(f) may be proved by Schwartz's
inequality.
The inequality is eviaentlj true ir
I I
(x,. -x)2 / ax I f(x)dx, but by Schwartz's C. -
d'-r
î i rfX212 inequality, the latter member is not smaller thanf dxj
L'j. 3
which is obviously (x2 - x1)2. Hence the inequality is
proved.
It can e&sily be shown that the general integral
inequality still holds a x increases without bound.
1k
V. SUMMATION BY HARMONIC MEANS
It has been shown that an infinite series may be
summed by the uso of arithmetic means of the partial suma. We shall now show that this may aleo be done by
harmonic means as well. Further, the method of harmonic
means is regular; that is, when applied to a convergent
series it will yield the ordinary sum.
Consider the series u where again we shell k=1
n let s = U. If the series is convergent, let
S = 11m s We must add the restriction that none of the n
partial sums nor S be zero. Then,
um i. 1, and it has been shown in section II that n-
lims1+e2+...+s S. 4CO
n
Let p I and P 1. Then lin p= P.
The p1 may now be considered as the partial sums of some
convergent series. Thus
11m p1 + p2 + ... + p P, or =
n
hm __ n n+ool+r ..
1 s2
Hence, if a series is convergent in the ordinary sense,
the limit of the harmonic mean of the partial sums is
15
equal to the ordinary sui». That i», the method is regular.
If, however, the series to be tested is not ordinarily
convergent but the limit of the harmonic mean exists, the
limit may be considered to be a conventional sum for the
series. We shall denote b,r H the lii» u i +.. i
s s I 2 n
when such exists. It follows from the definition that if the series u diverges to + or - 00 , so F1180 will
k=1
H. The most interesting application of this method, then,
will be to series whose partial sums oscillate between
finite or infinite limits. For an example, let us consider the series
F1 + I - I + I - I + ... Then 81=a
a + 1
83 = s.
a + i
s =a 2 n-1
52n =a+1. Nov, H21_ 2n-1 2n-1
- 1 +T. + i u + n - i ¡ a-+1 ¡
2n - I (2n - i)Ça)Sa + i), ___________ 2ia-+ i
a +1J and lia H 2a(a +
2a+1
-1 (;
Also, H2 2n__ 2a(i. 1 ñ+ n
Thus, the given serLes is swnblo by the harmonic mean
to the value 2e(a + i).
From this result we can see that the haraonic
mean sum of a nonconvergent series is not necessarily
linear. That t, chancing a by an amount b does not
necessarily change the hmonic mean sum by b.
Let us investigate the result of interpolating an
infinite number of zeros in the series of the previous
example. Corwider the series
a+1+O-i+i+O-1+1+O- Then s=a+1
83n+1. =
= a + j Now 3n 3a'a + i)
n 2h
And H3n+i = ! (n i- 1)ja1a + i) n+1+ 2n )na+n+a+1
a
Thus lUn H31 3a(a + i) 3a+I Aleo 11 2 = -
3n+ 2 on + 2 (a)[ + 1 + n +I+ 2ii-:i:--I 3naÏ- n+ ¿a + 1
-i:--- ----:i:--i
- 3aIa + i) and 1imH32
The harmonic mean sum is now 3a(a + i). Hence, the a +-I--
17
interpolation of an infinite number of zeros may affect
the sum. This is also true of Cesro summation. This
interpolation does not always affect the result, for an
examination of the series
a+1+O-1+O1+... will show the harmonic mean. sum still 2a(a + , as it
was without the zeros.
The typo of serios most frequently encountered in
summation by means is that in. which, after a finite (k)
number of terna, is recurrent with a fixed period (a)
of terms, and O < n K M, ni and M finite. Then
H n n
= .-. + ---+ . ..- +
For convenience, let (p1 + ... + and let
(pk1+...+pk+a)=A. ThenHk+ k+na fl8
= K+ii
and H k + na + d, where d is an integer such
that I d a, and D = p «. i- k +
D, then,
can take on at most a finite (a) number of values. Then
liuiHk+na=lifllHk+na+da
From the last result it is evident that the
restriction must be added that A O. Under these con-
ditions, then, a series is summable by the harmonic mean
method. It may easily be shown. that it is al8o summable
(c,i) under these conditions but not necessarily to the
same value, as we have seen in some of the previous
example s.
N0E: Suinmability by Cesro means does not neces-
sarily imply suimaability by iarmonic means, nor is the
reverse true. For example, consider the sequence
1' 2' 82n' 82n i. 1' ' where the pair 82n - i'
52n. is either 1, 1 or 1/10, 19/10. For n = 1, 2, ... (i) i , and
(i) = either I or 2m + 1/10
LU +
(i) Hence, liza C = I
Now, H _____________________ , n - I-:: I +-.; + I
'A n
+1 I 2 82n-1 82n,
2n
where the pair I , I is either 1, 1 or 10, 10
82n-1 82n
Then î. i.. th aritbmetic mean of n pairs which are 2n
either 1, 1 or 10, 10/19. The aritbnietic mean of 1, I
is 1, and that of IO, 10/19 is 100/19.
Construct si, 2' ' as follows: let the first
19
two terms be 1, 1. Then 1 1. Now use the pair 1/10.
19/10 repeated enough times so that the reciprocal of
the harmonic mean for these terms exceed8 k. Next, use
enough of the pairs 1, 1 80 that the reciprocal of the
harmonic mean is less than 2. When this process is con-
tinued, a series is obtained for which the Cesro sum is
1, but the harmonic mean oscillates, never approaching a
limit. Obviously, a similar method could be used to
derive a series summable by harmonic neans but not by
Cesro a1eeris.
It is nov possible to define higher orders of har-
monic means. A slight change in notation is necessary.
Let n . We have seen that i1' -
: ; i
1 8
-p , then -y 8. Then let
n n - + -F...+ 1?
Since 1H} is a sequence approaching the value S,
2i S. In general,
H_____n ______ r n -- rr. .
- r - r -
and rHn -, S
We may now state a definition of order. If r is
20
the smallest integer such that urn H exista, then n
u is summable (rH) by harmonic means of order r.
ki
From the above, it is apparent that if a series is sum-
mable (rH) it is sumniable (r +
1H) to the same value.
If u is convergent, it is sumrnable (rl!) for r k1
positive integer.
21
VI. TRE WEIGHTED HARMONIC MEAN
00 n
Cone ider the series . u, where s = u O. k=1 k=1
If the series is convergent, urn s S. Let I
Sn
Then P = urn p = urn I I . Consider n flPco
n
Cn , where
=a1+cn_1P2+ s..
n C = , and um O. Then
k=1 n-Ç +
en IP2 + +
w=____ --c-
and hm I = P I (by Norlund's Mean).
flc0
Thus, him = S, and this method is regular. As with n-400
the Närlund mean, if all c = 1, this method reduces to
the ordinary harmonic mean.
2
VII. SUMMATION BY GEOMETRIC MEAN
fl
Again consider 2 u1 where s = u s - O.
k=1. n k=1 k
We may define a geometric conventional sum as
follows: Let G (8152...8)hl'fl. Then
liai (2152.. e)h/'1
= G, the geometric conventional sum.
Phis method is regular, for by section III we saw that
H G A, and if I Uk is ordinari1 convergent to S,
k=1.
H A = S. Thus, G also equals 5. However, G may
exist when S does not. In that case G is a conventional
sum for the series.
There are certain restrictions that must be
placed upon the geometric meen. Clearly, in order to
have a value different from zero, none of the partial
sums can be zero. Also, in order to avoid imaginary
results we must insist that the product of the partial
sums be positive for al]. n N, where N is a finite
integer.
For an illustration of this method, let us con-
aider our familiar series, a + i - i + i - i + ...
82n = a + I
52n + I a
I/2n 1/2
G2n = [a(a + 1)nj
[a(a + i)J
2
I n+1 ri
rn + 1 )fl]2fl+'l ri+L
+ . = La (a + I a a
1)2flI
1/2
Then, urn = ' 2n + i = [a(a + 1)7
fl-CK
Thui, the serles te suiuinable by the geometric mean
1/2 to the valuo lIa(a + l)J
VIII TRE WEIGHTED GEOMETRIC MKAN
00
Once again consider the series where k=í
n = and ali are greeter then zero. Consider
k=1
i
s1)7vhere the e1 and C ere
as defined in section VII. Then
in .
- + .
+ e1 ina
nd 11m in G = lin in s = in S, provided this exists.
Therefore, G S if the &eries is convergent to k=1
S, and the method is regular.
X. DIVERGENT IN?INITE PRODTJCT
In considering infinite series we have worked with
conventional sums, regular methods of summation that
yielded the ordinary sum when applied to convergent
series but that also gave rnrnerical values wten apolied
to certain nonconvergent series. In a similar way we can
f md a method of evaluating infi1te products that is
regular; that is, Lb yields the ordinary product when
applied to a convergent infinite product but also may be
extended to assign a numerical value to certain noncon-
vergent products.
Consider the product TTa1 , and let p = a1a2...a.
Let G . Then we have
in G in p1 + in p2 + ... + in p
-- n
If e is convergent, thon the sequence (in )
converges to lnîíwherelTis the ordtnry value of the
infinite product. Hence, by arithmetic mean summation,
in G - in iT, or urn G =TF. However, if um p
does not exist, but um 0tl does exist, then urn 0ri may be n-ì°o
co
considered to be a conventional product of IT a,
Since this method is equivalent to that of Holder,
26
we may extend it easily to higher orders by taking the
limit of the root of the partial products of the partial
product s.
Let us apply this method to the infinite product
(1)(2)(1/2)(2)(1/2)... . Now
p1 =1
p2 =2 ........... 2n =2
2n + ± = 1
Obviously, the product does not converge. flowever,
1/2 n
= (i)(2)(1)(2)...(2)] 21/'2, and
I n
02n + = [(I)(2)(1)(2)...(2)(1)] =
Then, uni = lu a2 + 21/'2. Therefore, 21'2 ja
the conventional product.
L FURTHER EXAMPLE$ OF TILE &RMOtIC MEhN METHOD
Consider the sertes
U=1_ I I
In tht8 SetieE, .pplyirg,
first ten terms to obtain
0.685. The aun of
I I I I I I I + r - + 7 - + +
tc hernicnic ìieri method to the
&n approximation, we find that
the first ten terms is 0.61456.
r
The velue cf the infinite series is In 2 or 0.693. Hence,
the method of the haxinonic mean is in this case a better
approxiiwtion than is the partial sum.
Next, consider the series
for which
U=2-5-2+7-5-2-i-7-5-2+7-
=2
83
... . 0
s. 2 fl
+ I = 2
83n + 2
Here the conditions of the result in section V concerning
series with oscilleting partial sums are satisfied with
a = 3 and A = (-1/5 + 1/2 - 1/3) - 1/30. Hence, we
expect the result lin = )/(-I/30) = -90. In order to )-3X L
28
check this', we find tht
3n -90, -n + n - n
+ I = - + -90, anc
+ 2 __n +2 , 11m
+ 2 = -90.
-n+1-1 n-'°o 33 Thus, the series is summb1e (1H) to the value -90.
XI. COMPARISON OF CESARO AND HARMONIC SUMMATION
Consider the series U = a, where s k=1 k=1
and D (s/ M, M finite. If the series is con-
vergerit, let its sum be denoted by S. Let
(i) C I + W + n , and let 4H n We may
n = -- - - I. - i 3_. 3
J- n
now state the following properties or characteristics of
Cesàro and harmonic summation:
I. If U is convergent, it is suxnmeble (c,i) and () to the value S.
2. If U ta eummï.ble (C,r), it is eummble (C,r + i), and if it is aummable (rH) it is simmble r + 1H).
3. If U diverges to + or - 00 then both the
arithmetic and harmonic means diverge to +aO or .00
k. IC U is suinniable (c,i) or (1H), it either converges
or oscillates between finite or infinite limits.
(i) (i) 5. If fl C1 = 11m Cr = ' and finite, thon
U is said to be bounded (c,i). Similarly, if n1n=i;, I11n' 'nd$ftnite, then n- fl-3
U is said to be bounded (1H).
6. The interpolation of a finite number of zeros does
30
not chníe either the Cesro or harmonic $um.
'T. The interpolation of an infinite number of zeros
my chance both the Ces&ro and harmonic sums.
8. If U is summs.ble (C,r) and (.H), then rU is sum-
mable (C,r) and (rH) to r timos the former velues.
BIBLIOGRi.Pift
1. Cauchy, Augustin Loui3. OEvres coniplètes, series 2, vol. 3. Parie, Gauthier-Villare, 1897. 512p.
2. Hardy, G. H., Littlewood, J. E., and P6lya, G. In-
equalitiea. London, C8mbridge University press, 193k. Jlkp.
3. Hobson, E. W. The theory of functions of a real variable and the theory of Fourier's series, vol. 2. London, Cambridge University press, 1926. 78Op.
. Moore, Charles N. Summable sertes and convergence factors. New York, American Mathematical Society, 1938. lO5p. (American Mathematical Society colloquium publications, vol. 22)
5. Pólya, G. and Szegó', G. Aufgaben und lehrsitze aus der analysis. Berlin, Julius Springer, 1925. JJ8p.
6. Sznil, Lloyd L. History and synopsis of the theory of suznmable infinite processes 2:8. Eugene, University of Oregon publication, 1925. 175p.
7. Widder, David V. Advanced calculus. New York, Prentice-Hall, Inc., 197. k32p.