home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!wupost!cs.utexas.edu!sun-barr!ames!agate!linus!linus.mitre.org!gauss!bs
- From: bs@gauss.mitre.org (Robert D. Silverman)
- Subject: Re: ALGEBRAIC NUMBER ARITHMETIC
- Message-ID: <1992Jul30.163746.21362@linus.mitre.org>
- Keywords: algebraic numbers, arithmetic
- Sender: news@linus.mitre.org (News Service)
- Nntp-Posting-Host: gauss.mitre.org
- Organization: Research Computer Facility, MITRE Corporation, Bedford, MA
- References: <1992Jul30.151149.12688@ugle.unit.no>
- Date: Thu, 30 Jul 1992 16:37:46 GMT
- Lines: 92
-
- In article <1992Jul30.151149.12688@ugle.unit.no> ap@levangerhs.no (Andrei Prasolov) writes:
- >Answering to <aet.712408532@munagin>
- >in <1992Jul29.163830.6443@ugle.unit.no>
- >I made a mistake. If we represent algebraic numbers
- >only by their minimal polynomials, we cannot construct
- >a correct arithmetic. Let us take the main root x of
- ^^^^
-
- There is no such animal as a 'main root'. There is only a
- root and its conjugates. If alpha is a root of a polynomial,
- then the field Q(alpha) is isomorphic to Q(alpha') where
- alpha' is any of alpha's conjugates. The map is controlled by the
- Galois group.
-
- Your assertion that one cannot construct a correct arithmetic if all
- one has is the minimal polynomial is FALSE.
-
- > 2
- >f(X) = X -2 and the main root y of
- >
- > 4
- >g(Y) = Y -2. The algorythm says that z = x+y is
-
- The fields Q(x), Q(y) are different. The field Q(x+y) is an extension
- of Q(x) and also an extension of Q(y).
-
- >a root of h(Z) = h_1(Z) * h_2(Z) where
- >h_1(Z) = (Z^2+2)^2-2(2Z+1)^2
- > has roots x+y, x-y, -x+iy, -x-iy, and
- >h_2(Z) = (Z^2+2)^2-2(2Z-1)^2
- > has roots x+iy, x-iy, -x+y, -x-y.
-
- Which algorithm? I'm afraid I don't follow you when you say "the" algorithm.
-
- >The algorythm cannot determine, which polynomial
- >from the two the number x+y is a root of.
-
- So? This is totally irrelevent if working in Q(x+y). Your polynomial
- h(Z) is not irreducible. You further confuse the issues involved by writing
- terms like x + iy because the field Q(x) is purely real, while Q(y) is
- imaginary.
-
-
- I must confess that I don't understand what you are trying to say.
-
- > So the representation by pairs (irr. polynomial, interval)
- >seems to be reasonable. One needs, however, the following
- >two algorythms:
- > 1. Separation of roots of a polynomial
- > (there exists a lot of algorythms).
- > 2. Decomposition of polynomials over Q into
- > prime factors (I am not aware of any).
-
- This is totally irrelevent.
-
- Let me try to give a simple overview, along with an example.
-
- Let f(x) be the minimal (irreducible) polynomial of alpha. Assume
- f has degree d.
-
- Any element of the (algebraic extension) field Q(alpha) may be represented
- by a1 + a2 * alpha + a3* alpha^2 + .... a_d alpha^(d-1). This element is
- an algebraic number. I use the notation a = [a1,a2,...a_d] to represent
- one such number. Now, to perform arithmetic [let's say multiplication] on
- a and b [a*b = [a1,a2,...a_d] * [b1,b2... b_d] ] one does this as one would
- ordinary multiplication of two polynomials in alpha subject to the simplifying
- side condition f(alpha) = 0.
-
- For example. Consider Q(alpha = sqrt(2)) with f(x) = x^2 - 2.
- Now, (a1 + a2 alpha)*(b1 + b2 alpha) = a1 b1 + (a2 b1 + a1 b2) alpha + a2 b2 alpha^2.
-
- Use the fact that alpha^2 = 2 and this simplifies to:
-
- a1 b1 + 2 a2 b2 + (a2 b1 + a1 b2) alpha.
-
-
- Where's the problem???? Basic arithmetic in algebraic number fields [e.g.
- addition subtraction, multiplication] is trivial. Division is a little more
- complicated because one must find a representation for an inverse via (say)
- the extended Euclidean algorithm or by linear algebra.
-
- Now arithmetic in the ring of integers of an extension field Q(alpha) can
- get more complicated, because one must find an integral basis for the ring.
- It is currently an open field of study to find good algorithms for doing so
- when the field discriminant is large.
-
- I hope this clarifies things.
- --
- Bob Silverman
- These are my opinions and not MITRE's.
- Mitre Corporation, Bedford, MA 01730
- "You can lead a horse's ass to knowledge, but you can't make him think"
-