Special Issues of the Journal of Symbolic Computation
Report by Franz Winkler, RISC, Linz (Austria)
Guest editors:
- C.M. Hoffmann (Purdue Univ., USA)
- J.R. Sendra (Univ. Alcala, Madrid, Spain)
- F. Winkler (Univ. Linz, Austria)
A special issue of the JSC on "Parametric Algebraic Curves and
Applications" will appear soon, i.e. either in the last issue of
1996 or one of the first issues of 1997.
The main topics treated in this issue are implicitization and
parametrization problems and applications in computer aided geometric
design. According to this classification,
three different groups of papers can be distinguished.
The first group is concerned with implicitization of
parametrically presented curves and surfaces.
Starting from a parametrization of an algebraic curve or surface we
might be interested in deriving algebraic equations implicitly defining
the given geometric object. This is the problem of implicitization,
which amounts to a special problem in polynomial elimination theory.
Gonzalez-Vega, instead of using elimination methods such as resultants or
Gröbner bases, proposes to use certain symmetric functions,
Newton sums, to solve these special elimination problems.
Sederberg, Goldman, and Du propose the use of moving algebraic curves
for implicitizing algebraic curves. In the absence of base points,
moving lines turn out to be sufficient for implicitizing rational curves.
Circular curves are traced by a point on a circle rotating around
a center, which again is a point on a circle rotating aroung a center,
which again ... . Such circular curves can be given parametrically
in terms of sines and cosines. Hong presents an
implicitization algorithm for such parametrically given curves.
The second group of papers is concerned with the inverse problem, namely
parametrization of implicitly given algebraic curves and surfaces, as
well as reparametrization.
Sendra and Winkler continue their work on deriving a parametrization
algorithm for algebraic curves. Whenever an algebraic curve can be parametrized,
then there is a variety of different parametrizations.
So one wants to look for a parametrization with good properties, such
as being defined with only rational or real coefficients.
A theorem by Hilbert and Hurwitz is generalized, which allows to reduce
such optimality questions to investigations of conics.
Alternatively, van Hoeij presents an algebraic approach based on canonical
divisors. From a basis of
the linear space
of a rational divisor
on a curve
, one gets a bijective morphism from
to a conic. So questions such as finding points or parametrizations
over low degree algebraic extensions can be investigated for the conic
and transferred to the curve
.
Adjoint curves play an important rôle in the geometry of
algebraic curves, in particular in parametrization.
They are curves passing with certain multiplicity through the
singularities of an algebraic curve. From this geometric
specification one can derive linear conditions on the coefficients
of the adjoints. The problem is that also neighboring singularities
have to be considered, which means application of quadratic
transforms blowing up the degree. Mnuk analyzes the problem of
determining adjoints and describes a purely algebraic approach.
Real varities are the most common ones in geometric design. However,
since algebraic geometry is mainly developed over algebraically
closed fields, often reality questions are not treated.
Recio and Sendra investigate the problem of deciding whether a
parametrized complex curve is real, and if so how
to reparametrize it over the reals.
Finally, in this group of papers we have two contributions that analyze the
rationality and provide parametrization algorithms for special algebraic
varieties such as canal surfaces or generalized offsets.
Consider a space curve parametrically given by
, and a radius
function
. A canal surface is the envelope of the spheres with
center at
and radius
. In this situation Peternell and Pottmann
show that canal surfaces are rationally parametrizable, and a parametrization
algorithm from them is presented.
The generalized offset to a hypersurface
at distance
is
essentially the Zariski closure of the set of intersection points of
the spheres with center on
and radius
, and the lines
obtained when applying a direct isometry to the normal lines to
at
the center of the spheres. Arrondo, Sendra and Sendra characterize the
unirationality of the components of the generalized offsets to hypersurfaces,
and provide algorithms for computing rational parametrizations when these
components are parametric.
The last group of papers treats more applied problems in the area
of computer aided geometric design, such as constraint solvers or numerical
approaches.
The goal in the paper by Hoffmann and Joan-Arinyo
is to enhance constructive geometric
constrain solvers with the capability of managing functional relationships
between dimension variables. A reduction on clusters of geometric elements
is introduced so that geometric constraints can be propagated between
such clusters, and ultimately satisfied. Termination of this reduction
system is established, and if the problem is not geometrically
overconstrained then reduction leads to unique normal forms.
Bézier curves are parametric curves widely used in computer aided geometric
design. Farouki investigates conic approximations to offsets of
conic Bézier curves.
Exact upper bounds on the error incurred in such approximations are
derived.
Symbolic and numerical techniques are combined by Bajaj and Xu
to construct spline approximations of real algebraic surfaces.
This is achieved by a novel triangular expansion scheme on the surface
conforming to point and curve singularities on the surface.
Special issues of the JSC
SAC Newsletter's Home page