home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math.symbolic
- Path: sparky!uunet!munnari.oz.au!cs.mu.OZ.AU!eric!ross
- From: ross@ecr.mu.oz.au (Ross McAree)
- Subject: Help using Macaulay to convert the tangential equation of a surface into a point equation
- Message-ID: <9223113.10992@mulga.cs.mu.OZ.AU>
- Sender: news@cs.mu.OZ.AU
- Organization: Computer Science, University of Melbourne, Australia
- Date: Tue, 18 Aug 1992 03:14:15 GMT
- Lines: 72
-
-
-
- This problem described below is one that I've been struggling with
- for quite a while using maple. Unfortunately I haven't been able
- to get a result. A couple of people suggested that Macaulay's Grobner
- basis computations are most likely to yield a result. This being
- so I ftp'd Macaulay, however I find the language of Macaulay
- far too mathematical for a mechanical engineer. Can anyone show
- me that way clear through all these "rings" and "ideals" and so
- forth to a possible solution.
-
- The problem is one of converting a quartic envelop surface (QES) in R^3
- into a point surface equation. I suspect that the point surface is a
- sextic---certainly a number of sections through it are sextics. The
- equation of the QES in terms of tangent plane variables (t, u, v, s)
- is
-
- QES = (a^2 t^4 + b^2 u^4 + c^2 v^4)
- -2 {b c u^2 v^2 + c a v^2 t^2 + a b t^2 u^2 + 2(t^2 + u^2 + v^2)} = 0
-
- where (t, u, v) is the normal at each plane and s
- is the shortest distance of the plane from the origin. The three
- "coefficients" a, b, and c are all real and satisfy
-
- a + b + c = 0
-
- In finding a solution to the point equation it would not really
- bother me if I had to substitute in values for a, b, and c,
- though I'd prefer to do it symbolically.
-
- The "traditional" approach this problem is as follows. If the equation
- of the tangent envelope to the surface is
-
- (*) F(t, u, v, s) = 0.
-
- Then the partial derivatives Ft = dF/dt, Fu = dF/du, Fv = dF/dv,
- and Fs = dF/ds, define the homogeneous point (x, y, z, w):
-
- x = Ft,
- y = Fu,
- z = Fv,
- w = Fs,
-
- polar to the general plane (t, u, v, s). If this plane is tangential to
- the surface defined by (*), the polar point (x, y, z, w) is
- on the surface. The equation set defined by the four equations above
- and a further condition (namely the requirement that the point (x, y, z, w)
- be on the plane (t, u, v, s)), i.e
-
- (**) x * t + y * u + z * v + w * s = 0
-
- can be used to find the point equation of the surface by eliminating
- (t, u, v, s). Alternatively the function F could be used instead
- of equation (**) by virtue of Euler's formula for homogeneous
- polynomials.
-
- Unfortunately the elimination of t, u, v, s from the equation
- set above has me stumped. It also has maple stumped--I set it
- loose on the problem for two weeks on a sun 4/490
- and it didn't get anywhere). Surely this is not a difficult
- problem to solve and perhaps someone sufficiently familiar with
- Macaulay could point out how to set up the input file for
- this problem. I believe the script file dual_variety is the likely
- to be required here however I'm totally confused by the
- language of Algebraic geometry and I don't know where to
- begin.
-
- Ross McAree
- Robotics Laboratory,
- Department of Mechanical Engineering.
- The University of Melbourne.
- Parkville.
-