19.07.2014 Views

Contents - Student subdomain for University of Bath

Contents - Student subdomain for University of Bath

Contents - Student subdomain for University of Bath

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.

114 CHAPTER 4. MODULAR METHODS<br />

as being in R and check that it is indeed the right result. The key questions are<br />

then the following.<br />

1. Are there domains R i <strong>for</strong> which the behaviour <strong>of</strong> the computation in R i<br />

is sufficiently close to that in R <strong>for</strong> us to make deductions about the computation<br />

in R from that in R i ? Such R i will generally be called “good”.<br />

2. Can we tell, either immediately or with hindsight, whether an R i is<br />

“good”? It will <strong>of</strong>ten turn out that we can’t, but may be able to say<br />

that, given R i and R j , one <strong>of</strong> them is definitely “bad”, and preferably<br />

which.<br />

3. How many reductions should we take? In practice we will only count<br />

“good” reductions, so this question is bound up with the previous one.<br />

4. How do we combine the results from the various R i ? The answer to this<br />

will <strong>of</strong>ten turn out to be a variant <strong>of</strong> the Chinese Remainder Theorem.<br />

5. How do we check the result? Can we be absolutely certain that this result<br />

is in fact the answer to the original question? In category speak, does<br />

Figure 4.1 commute?<br />

4.1 Gcd in one variable<br />

Let us consider Brown’s example (from which page 46 was modified):<br />

A(x) = x 8 + x 6 − 3x 4 − 3x 3 + 8x 2 + 2x − 5; (4.1)<br />

B(x) = 3x 6 + 5x 4 − 4x 2 − 9x + 21. (4.2)<br />

Let us suppose that these two polynomials have a common factor, that is a<br />

polynomial P (<strong>of</strong> non-zero degree) which divides A and B. Then there is a<br />

polynomial Q such that A = P Q. This equation still holds if we take each<br />

coefficient as an integer modulo 5. If we write P 5 to signify the polynomial P<br />

considered as a polynomial with coefficients modulo 5, this equation implies that<br />

P 5 divides A 5 . Similarly, P 5 divides B 5 , and there<strong>for</strong>e it is a common factor 1 <strong>of</strong><br />

A 5 and B 5 . But calculating the g.c.d. <strong>of</strong> A 5 and B 5 is fairly easy:<br />

A 5 (x) = x 8 + x 6 + 2x 4 + 2x 3 + 3x 2 + 2x;<br />

B 5 (x) = 3x 6 + x 2 + x + 1;<br />

C 5 (x) = remainder(A 5 (x), B 5 (x)) = A 5 (x) + 3(x 2 + 1)B 5 (x) = 4x 2 + 3;<br />

D 5 (x) = remainder(B 5 (x), C 5 (x)) = B 5 (x) + (x 4 + 4x 2 + 3)C 5 (x) = x;<br />

E 5 (x) = remainder(C 5 (x), D 5 (x)) = C 5 (x) + xD 5 (x) = 3.<br />

Thus A 5 and B 5 are relatively prime, which implies that P 5 = 1. As the leading<br />

coefficient <strong>of</strong> P has to be one, we deduce that P = 1.<br />

1 Note that we cannot deduce that P 5 = gcd(A 5 , B 5 ): a counter-example is A = x − 3,<br />

B = x + 2, where P = 1, but A 5 = B 5 = x + 2, and so gcd(A 5 , B 5 ) = x + 2, whereas P 5 = 1.

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

Saved successfully!

Ooh no, something went wrong!