home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!europa.asd.contel.com!darwin.sura.net!jvnc.net!netnews.upenn.edu!netnews.cc.lehigh.edu!ns1.cc.lehigh.edu!fc03
- From: fc03@ns1.cc.lehigh.edu (Frederick W. Chapman)
- Newsgroups: sci.math
- Subject: Re: polynomial roots and zero divisors
- Message-ID: <1992Aug28.152529.113932@ns1.cc.lehigh.edu>
- Date: 28 Aug 92 15:25:29 GMT
- Organization: Lehigh University
- Lines: 96
-
- In article <27AUG199211450866@cs.umass.edu>, rcollins@cs.umass.edu
- (Bob Collins) writes:
-
- > I would like to find the eigenvalues/vectors of a square
- >matrix. Unfortunately, the elements of the matrix are not
- >members of a field. In particular, they are "dual numbers",
- >defined in Yaglom's "Complex Numbers in Geometry" as a type of
- >complex number (a + b E), a and b being reals, and E being a
- >nilpotent imaginary unit; so E^2 = 0. Dual numbers form a
- >commutative ring, but they are not a field since any number
- >of the form (0 + a E) is a zero divisor.
- >
- >Question: Can I find the eigenvalues of a matrix (roots of a
- >characteristic polynomial) whose elements (coefficients) are
- >from a commutative ring, but not a field?
-
- [...]
-
- Certainly! In this case, it's not really any more difficult than solving
- individual polynomial equations over the reals. Suppose you want to solve
- the following quadratic equation over the "dual numbers":
-
- (x + y E)^2 + (-3 + E)(x + y E) + (2 + 4 E) = (0 + 0 E).
-
- Expand all the multiplications and simplify using the rule E^2 = 0:
-
- x^2 + 2xy E - 3x + x E - 3y E + 2 + 4 E = 0 + 0 E.
-
- Now collect terms together:
-
- (x^2 - 3x + 2) + (2xy + x - 3y + 4) E = 0 + 0 E.
-
- Equate the coefficients of the two components to get a system of real
- polynomial equations in x and y:
-
- x^2 - 3x + 2 = 0
- (2x - 3)y + (x + 4) = 0.
-
- The first equation involves only x; solve the first equation to get
-
- x = 1 or x = 2.
-
- Substitute x = 1 into the second equation and solve for y to get y = 5.
- Substitute x = 2 into the second equation and solve for y to get y = -6.
- Thus, 1 + 5 E and 2 - 6 E are the only solutions to the original quadratic
- equation over the dual numbers.
-
- ...........................................................................
-
- In general, since
-
- (x + y E)^n = x^n + (n x^{n-1} y) E
-
- and
-
- (a + b E)(x + y E)^n = a x^n + (a n x^{n-1} y + b x^n ) E,
-
- we can conclude that any polynomial equation in the unknown x + y E over
- the dual numbers will always simplify into a system of two real polynomial
- equations in x and y; the first equation will always be a polynomial in x
- alone, and the second equation will always be of degree one in y, as
- follows:
-
- p(x) = 0
- p'(x) y + q(x) = 0
-
- where p'(x) denotes the derivative of the polynomial p(x). Solve the first
- equation for REAL solutions x = r, using a suitable numerical method if
- necessary. (To make use of COMPLEX solutions x = u + v I, we must extend
- the "REAL dual numbers" to form the "COMPLEX dual numbers"; i.e., numbers
- of the form (a + b I) + (c + d I) E, where I^2 = -1, E^2 = 0, IE = EI, and
- a, b, c, and d are real; you may or may not wish to make this extension.)
-
- Upon substituting x = r into the second equation, we find that there are
- three things that can happen. (1) If r is NOT a multiple root of p(x) --
- that is, if p'(r) is non-zero -- then the second equation has a unique
- solution y = -q(r)/p'(r), and r - q(r)/p'(r) E is a solution to the
- original polynomial equation over the dual numbers. (2) If r IS a multiple
- root of p(x) -- that is, if p'(r) = 0 -- but r is NOT a root of q(x), then
- the second equation has no solutions; thus, there is no real y for which
- r + y E is a solution of the original polynomial equation over the dual
- numbers. (3) If r IS a multiple root of p(x) AND r is a root of q(x), then
- all real y satisfy the second equation, thereby yielding an infinite family
- of solutions r + y E to the original polynomial equation over the dual
- numbers).
-
- Tha tha tha tha tha tha tha tha that's all folks! :-)
-
- --
-
- o ------------------------------------------------------------------------- o
- | Frederick W. Chapman, User Services, Computing Center, Lehigh University |
- | Campus Phone: 8-3218 Preferred E-mail Address: fc03@Lehigh.Edu |
- | "The day after yesterday is the second-to-last day before |
- | the rest of your life the day after tomorrow." |
- o ------------------------------------------------------------------------- o
-