[Edu-sig] "The study of fixed points has been at the foundation of algorithms"
Arthur
ajsiegel at optonline.net
Wed Dec 14 15:47:46 CET 2005
A bit of a windy road:
starting, as usual, with the personal frame of reference....
PyGeo's current implementation supports the exploration of the geometry
of complex numbers, and therefore speaks Mobius transformations.
http://pygeo.sourceforge.net
now has a pretty picture of a simple recursive transformation of 4
circles on the unit sphere (...thanks to __iter__ the ability to
recursively transform any arbitrary set of geometric objects is now
built-in to PyGeo).
My current exploration (current as in today) is finding the mechanism to
build a Mobius transformation that would be based (in part) on it's
(pickable and movable) fixed points - of which a Mobius transformation
has 2, which may coincide, or be located inconveniently - e.g. at infinity.
Which has me stepping into the math of the fixed points of a function -
it being trivial to find the fixed points, given the Mobius
transformation matrix, but less trivial (from where I am sitting at the
moment) to build the transformation from fixed point information. So I
am struggling and researching some.
In the course of which I come across this definition of "Fixed Point"
in a programming glossary, @
http://carbon.cudenver.edu/~hgreenbe/glossary/second.php?page=F.html
Of a function, f:X-->X, f(x)=x. Of a point-to-set map, F:X-->2^X, x is
in F(x). The study of fixed points has been at the foundation of
algorithms
<http://carbon.cudenver.edu/%7Ehgreenbe/glossary/second.php?page=A.html#Algorithm>.
Having discussed here my growing interest in some study of algorithmics,
but not getting there yet, but pursuing something that as far as I am
aware is unconnected to such study, and then finding this statement
indicating there is more of a connection - perhaps - than I had
understood, is to me interesting.
I have thought of "fixed point" (in programming) as connected
to/opposed to "floating point", not as something directly connected to
the concept of "f(x)=x"
The statement above seems to be telling me otherwise.
Guess I am fishing for some exposition on the statement that the
"The study of fixed points has been at the foundation of algorithms"
Art
More information about the Edu-sig
mailing list