03.01.2015 Views

Vol. 10 No 7 - Pi Mu Epsilon

Vol. 10 No 7 - Pi Mu Epsilon

Vol. 10 No 7 - Pi Mu Epsilon

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

550 PI MU EPSILON JOURNAL<br />

By proceeding in the manner described in the preceding paragraph, the<br />

follo\ving rearrangement of the given series is obtained:<br />

1 1 1 l<br />

-1 (l '0)-2 ( 1. 0)-3 (l' 0)-4 ( 1 '0)- 1 ( -1.-1) -2 (-I' -1)- 1(0, 1)<br />

- _!_( l , 0) - _!_( 1, 0) - _!_( 1 , 0) - _!_( 1, 0) - _!_( 1 , 0) - _I (1 , 0) - - 1 ( 1, 0)<br />

5 6 7 8 9 <strong>10</strong> 11<br />

l 1 l l 1 1<br />

-3( -1, -1)- 4( -1, -1)- 2(0, 1) -12(1,0) -13(1,0) -14(1,0)<br />

1 I l I l<br />

-15(1,0)-16(1,0) -17(1,0) -18(1,0) -19(1,0)- ···,<br />

which appears to be converging to S = (-1, 1). Levy then uses a complicated<br />

geometric approach to show that the rearranged series actually converges to S<br />

= (-1, 1).<br />

In the last paragraph of his paper, Levy claims that his method may be<br />

generalized to consider rearrangements of infinite series of vectors in n­<br />

dimensional space [6, p. 511]. His main idea is to decompose the series under<br />

consideration into n + 1 partial series. This corresponds to the idea of<br />

decomposing a series of complex numbers, or equivalently a series in R 2 , into<br />

three partial series. Levy does not, however, elaborate on exactly how this<br />

procedure would work.<br />

Levy's article is very concise. It offers neither proofs nor examples, and is<br />

somewhat poorly written. Knopp [5, p. 398] and Kadets [4, p. 1] state that Levy<br />

proved the result on the description of the set of all rearrangements of a series of<br />

complex numbers but they make no mention of the fact that he at least<br />

conjectured the result for n-dimensional space. Rosenthal claims [9, p. 342] that<br />

it was Levy who first proved the result for n-dimensional space. The fact is that<br />

Levy did not prove the result, but he was at least aware of it.<br />

III. Steinitz' Results.<br />

Steinitz begins the introduction to his paper on conditionally convergent<br />

series in convex systems by briefly tracing the history of the study of the<br />

rearrangement of series of real numbers. He introduces the term "summation<br />

range" of a series.<br />

RANGEMENT OF SERIES, SMITH 551<br />

finition 1. Let X be a finite dimensional vector space. The summation<br />

00<br />

of the series L ~ where ~ E X, is the set of all vectors x E X such that<br />

k=l<br />

L "-..(k> converges to x where L "-n(k> is a rearrangement of L xk.<br />

t:l k=l k=l<br />

Steinitz then restates Riemann's result as follows: The summation range of<br />

conditionally convergent series of real numbers is the set of all real numbers<br />

[<strong>10</strong>, p. 128]. Next he considers the summation range of a complex series ~nd<br />

gJ\'es two brief examples in the complex plane [<strong>10</strong>, p. 128-129]. The followmg<br />

example, Steinitz first, demonstrates that the summation range of a complex<br />

series may be a line in the complex plane.<br />

Let g be any line in the plane, a a point on g, and w =I= (0, 0) any point<br />

co<br />

on the line through the origin parallel to g. Let L 3n be any conditionally<br />

n=l<br />

convergent series of real numbers. Using standard vector addition it is easy to<br />

see that any number of the form a + rw, where r is a real number, will be a point<br />

00<br />

on the line g. By Riemann's theorem, the series L 3n can take on any real<br />

n=l<br />

value r. Hence the complex series a + a 1 w + ~ w + ··· can take on every complex<br />

value of the form a +rw. Thus the line g is the summation range of the complex<br />

senes.<br />

Steinitz second example demonstrates that a series of complex numbers can<br />

be rearranged to converge to any point in the complex plane. He begins<br />

00 00<br />

\\ith two conditionally convergent series of real numbers L 3n and L bn.<br />

00<br />

Again using Riemann's theorem, L 3n can be rearranged to converge to any<br />

n=l<br />

00<br />

real number a, and L bn can be rearranged to converge to any real number b.<br />

n=l<br />

Thus the series formed from the numbers a 1 , b 1 i, ~' b 2 i, ···can be rearranged .. to<br />

converge to any point in the complex plane.<br />

Steinitz makes two assertions in his introduction: first, that the summation<br />

range of a conditionally convergent series of complex numbers will always be<br />

n=l<br />

n=l

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!