Re: Generation of microstructure

"Marine Tort" <marine.tort@xxxxxxxxxxxxx> wrote in message
<g0r8a2$58v$1@xxxxxxxxxxxxxxxxxx>...
Hi,
I have to generate the geometry of a micro structure using
Matlab which will be incorporate as an input file in a
modeling software.

I have a problem with the generation of such geometry :
the
method I have to use is Sphere Growth Algorithm : the
cells
are circles (or spheres) randomly allocated in space with
different sizes. The size of the circles is defined
accordingly to the desirable distribution.
The next step of the algorithm is the growth of the
circles
until they fill the space totally.

I don't manage to define the position of the circles
without
intersecting themselves and I don't see how I can obtain
the
last step.
Could anyone help me ?
Thanks a lot.

This entire problem looks it will require an iteresting
algorithm.

However, in reference to the part of the problem your
question is about....

Are you constrained on how large the domain is of the
sphere's initial positions? If not, perhaps you can start
with your set of initial positions and initial radii.
Then, determine the largest amount of overlap, and "expand"
the initial positions so that there is no overlap.

In fact, you can simply determine the sum of the 2 largest
radii (call this "sumR"), and the distance between the 2
closest centers (call this "minD"). Then, you can multiply
each x, y, and z value of the initial sphere centers
by "sumR-minD" (which will be the largest amount of overlap
possible). This method will guarantee that no 2 spheres
overlap.

If you would like more help on this, please reply. I would
like to play with it some more.
.

Relevant Pages

  • Re: The sphere and hyperbolic geometry
    ... My geometry - the surface of the sphere, all points and all circles ... axiom, ...
    (sci.math)
  • Re: The sphere and hyperbolic geometry
    ... My geometry - the surface of the sphere, all points and all circles ... axiom, ...
    (sci.math)
  • Re: Curved space
    ... Do we have a material difference of opinion on great circles and arcs? ... Surely we can agree that on a sphere - which the Earth is not, ... But Euclidean geometry is merely the outworking of certain ...
    (talk.origins)
  • a question about geometries
    ... Does the concept of geometry, which includes lines and points, require ... between any two points there exists what you're defining as a line, ... So, for example, if you try to define a geometry on a sphere ... sphere with all great circles as lines certainly does ...
    (sci.math)
  • Re: A spiral space-filling curve as a natural continuum
    ... geometry from other definitions, towards that then the idea is to get ... There is a notion that a spiral space-filling curve starts from the ... disc to begin and then to generalize) and R^N. ... Then, in considering the points of these circles, which have ...
    (sci.math)