Services & Resources / Wolfram Forums / MathGroup Archive
-----

MathGroup Archive 2014

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Function parameterization guess

  • To: mathgroup at smc.vnet.net
  • Subject: [mg132648] Re: Function parameterization guess
  • From: "djmpark" <djmpark at comcast.net>
  • Date: Sun, 27 Apr 2014 21:45:24 -0400 (EDT)
  • Delivered-to: l-mathgroup@mail-archive0.wolfram.com
  • Delivered-to: l-mathgroup@wolfram.com
  • Delivered-to: mathgroup-outx@smc.vnet.net
  • Delivered-to: mathgroup-newsendx@smc.vnet.net
  • References: <8720827.39330.1398579243598.JavaMail.root@m03>

Narasimham,

I'm rather wading into the edge of my knowledge, but I think that in general
the answer is no, or it is often difficult to find parametrizations.

If you want algebraic or rational parameterizations of algebraic equation
solutions then you can't always do it.

If you want to do calculus on the solution set then you can make it into a
manifold if the Jacobian is nowhere zero on the solution set. But then it
will often require a number of charts in an atlas that will cover the
manifold and this can be done in an infinite number of ways. For example,
the {Cos[t], Sin[t}} parameterization is not adequate because it is not a
1-1 map of R^1 to the circle. t = 0 and t = 2 Pi give the same point on the
circle and that ruins the calculus. You would need two overlapping angular
domains, or four projections to the axes, or two stereographic projections.

So generally it's a difficult or at least a nitty-gritty problem.


David Park
djmpark at comcast.net 
http://home.comcast.net/~djmpark/index.html 

 

From: Narasimham [mailto:mathma18 at gmail.com] 


Using excellent function capabilities of Mathematica is it not possible to
generally guess or propose some standard parameterizations of components
given functions?
 
For two variables and single parameter. Given x^2 + y^2 =1 we have {x,y}=
{Cos[t],Sin[t]} and its variants {Sech[t],Tanh[t]}among others are
solutions.

For three variables and two parameters. Given x^2 + y^2 - z^2 =1 we have
Cosh[u] Cos[v], Cosh[u] Sin[v], Sinh[u] and variants..

The number of parametric set variations for component variables is not
infinite, can be indicated with an arbitrary constant. A general or possible
sub parameterization may be considered for each functional relationship.

Regards
Narasimham




  • Prev by Date: Re: Better way to test the number of arguments?
  • Next by Date: Re: Better way to test the number of arguments?
  • Previous by thread: Function parameterization guess
  • Next by thread: missing in Manipulate