05.01.2013 Views

COMPUTATIONAL PROBLEMS IN ABSTRACT ALGEBRA.

COMPUTATIONAL PROBLEMS IN ABSTRACT ALGEBRA.

COMPUTATIONAL PROBLEMS IN ABSTRACT ALGEBRA.

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.

368 Shen Lin<br />

sequence of squares, the a,‘s settle down to around 5. Since we know that<br />

all sufficiently large odd numbers can be expressed as a sum of three<br />

or fewer primes and all sufficiently large numbers can be expressed<br />

as a sum of five or fewer distinct squares, we are led to the following conjecture<br />

and theorem :<br />

CONJECTURE. Lim sup a,, exists for all complete polynomial sequences<br />

and the sequence of primes, sequence of squares of primes, etc. In particular,<br />

[lim sup a,] = 3 for the sequence of primes and [lim sup a,] = 5 for the<br />

sequence generated by x2.<br />

Note that if this conjecture is true, then the status of the Goldbach<br />

conjecture can be settled by finite enumeration as can be seen from the<br />

following theorem :<br />

THEOREM: Let S E (~1, s2, . . . , Sk, . . .> be an essentially complete<br />

sequence. Suppose there exists an N and an a such that for all n - N,<br />

Wn,,)<br />

a, E ___<br />

%<br />

-C a. Then all sufficiently large integers can be expressed<br />

as a sum of at most [a] distinct terms from S, where [x] stands for the<br />

largest integer G x. As a consequence thereof, S forms an integral basis<br />

for large numbers of order at most [a]. At any rate, the conclusion is true<br />

for all numbers y, O(S,) -C y -z q(S,,+,), for which a, -K a.<br />

Proof. Let y =- 6(&J; then we may find an n Z= N such that 6(S,J -C y <<br />

e(S,+,). Since y is greater than 0(&J, y is representable as a sum of distinct<br />

terms from S,,, say y = Si,+Si, + . . . +Si, where each s4 3 s,. Hence<br />

y * ts,. On the other hand, y 4 0(S,,+,) < asn. Hence t -K a. Since t is<br />

an integer, t < [a].<br />

While computer work cannot yet establish the validity of the assumption<br />

needed in the above theorem, we believe that it can give us a fairly good<br />

indication of what a may be, if it exists. It is hoped that experimental<br />

work of this kind can help us formulate meaningful conjectures that some<br />

one can prove at a later date.<br />

APPENDIX A<br />

Thresholds of completeness have been computed for many sequences,<br />

and the behavior of the respective a,‘s studied. In the following tables we<br />

give a summary of the computed results obtained for some selected sequences.<br />

The generating function for each sequence (si = f(i)) is given on<br />

top of each table and max (an) means the largest a,, in the range between the<br />

two values of n heading the column in which the value of max(aJ is found.<br />

Sequences which form integral bases<br />

TABLE 1 f(x) = x2+1<br />

n 1 2 3 4 5 6 100 200 300 400<br />

&a 2 5 10 17 26 37 10,001 40,001 90,001 160,001<br />

WI) 51 131 255 282 360 465 54,916 196,116 415,347 726,436<br />

q&l<br />

a.-1 = - 65.5 51.0 28.2 21.2 17.7 5.602 4.952 4.645 4.562<br />

h-1<br />

max@,) 5.673 5.328 4.880<br />

TABLE 2 f(x) = x2<br />

n 1 2 34 5 6 50 100 150 200 250 350<br />

$98 1 4 9 16 25 36 2500 10,000 22,500 40,000 62,500 122,500<br />

x9(&) 128 192 223 384 492 636 17,072 60,928 129,184 222,208 339,968 659,456<br />

n<br />

&I<br />

1<br />

2<br />

&S,> 6<br />

~(S,><br />

asel = -I_<br />

G-1<br />

max@,)<br />

max(4<br />

369<br />

7.110 6.216 5.818 5.611 5.483 5.414<br />

7.110 6.346 5.893<br />

TABLE 3 P,, FE the sequence of primes<br />

2 3 4 5 100 500<br />

3 5 7 11 541 3571<br />

9 27 45 45 1683 10,779<br />

3.217 3.028<br />

5.729 5.621<br />

1000 2000<br />

7919 17,389<br />

23,859 52,247<br />

3.017 3.004<br />

3.217 3.044 3.032<br />

TABLE 4 Q,, E sequence of pseudo-primes<br />

2 3 4 5 500 1000 2000 3000<br />

3 4 5 6 1082 2307 4891 7619<br />

2 8 8 12 2172 4625 9835 15,257<br />

2.009 2.005 2.011 2.003<br />

2.049 2.037 2.025

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

Saved successfully!

Ooh no, something went wrong!