Re: Optimum spline problem

That helps. Thank you very much.

In article <d4l0bl$spb$1@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
spellucci@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx (Peter Spellucci) wrote:
>
>In article <bzibe.21122$JB.16908@xxxxxxxxxxxxxxxxxxxx>,
> John_W_Herman@xxxxxxxxx (John Herman) writes:
> >I'm trying to guess a good shape based on very sparce information. I thought
>
> >maybe there might be an optimization approach might allow for an improvement
> >in a shape over the "tinker-toy" approach I'm currently using.
>
>
>if you want to design a curve, much depends on the nature of points.
>normally, if you deal with a curve (not a function graph)
>you might parametrize with the distances of the interpolating
>piecewise linear arc and then interpolate by an ordinary cubic spline or
>a spline under tension. selection of the tension parameter is
>somewhat arbitrary. minimizing the total curvature makes it a nonlinear
>problem . hence anything depends on what you feel is a "nice curve"
>taking distances to the end points into account would yield something of
> minimal
>length, a bit similar to the spline under tension?
>hth
>peter
>
>
>
> >
> >
> >In article <d4iief$ifo$1@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
> >spellucci@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx (Peter Spellucci) wrote:
> >>
> >>In article <8tvae.3161$JB.2248@xxxxxxxxxxxxxxxxxxxx>,
> >> John_W_Herman@xxxxxxxxx (John Herman) writes:
> >> >I have a problem where I am trying to estimate the optimum shape of a
> line.
> >> I
> >> >estimates of the radial distance from the end node to each of the other
> nodes
> >>
> >> >and the tangent to the the curve at each point except the end node. The
> >> nodes
> >> >are irregularily spaced but the maximum distance between nodes and the
> >> maximum
> >> >distance from the end node are known exactly. What approach is best to
> use
> >> >for this problem?
> >>
> >>it is completely unclear to me what you want to "optimize".
> >>the distance to the endpoint: which role should it play here
> >>hth
> >>peter
.

Relevant Pages

  • Re: best fit curve
    ... Other than drawing by hand a nice smooth curve through the points and then resampling it can anyone suggest a better method. ... Is there a way of deciding what the optimum interval is between the spline knots to minimize the number of data points required to approximate the curve to the required accuracy? ... of course there exist codes for monotonic splines: ... has a spline under tension which can also avoid oscillations ...
    (sci.math.num-analysis)
  • Re: Optimum spline problem
    ... if you want to design a curve, much depends on the nature of points. ... a spline under tension. ... >>>estimates of the radial distance from the end node to each of the other nodes ...
    (sci.math.num-analysis)
Loading